Download presentation
Presentation is loading. Please wait.
1
Introduction To Business Process Design
Chapter 1 Business Process Modeling, Simulation and Design
2
Overview What is a business process?
Three definitions Process types and hierarchies Components of process architectures The essence of Business Process Design (BPD) Why is BPD important? BPD and overall business performance BPD and strategy Why do inefficient processes exist? Activity classification and BPD
3
What is a Business Process?
A pragmatic definition A Business Process describes how something is done in an organization In general terms… Business - Organizational entity that deploys resources to provide customers with desired products and services Process (Merriam-Webster’s Dictionary) (i) A natural phenomenon marked by gradual changes that lead to a particular result (ii) A natural continuing activity or function (iii) A series of actions and operations conducing to an end
4
What is a Business Process?
2. Traditional Process definition in OM literature A process specifies the transformation of inputs to outputs The transformation model of a process Inputs Outputs Process Different types of transformations Physical (Ex. raw material finished product) Locational (Ex. flying from Denver to L.A.) Transactional (Ex. depositing money in a bank) Informational (Ex. accounting data financial statement)
5
What is a Business Process?
The Process View Any organization entity or business can be characterized as a process or a network of processes Based on the simple transformation model of a process Has its origin in the areas of manufacturing and quality The transformation model of a process Inputs Outputs Process
6
What is a Business Process?
A more comprehensive process definition A business process is a network of connected activities and buffers with well defined boundaries and precedence relationships, which utilize resources to transform inputs into outputs with the purpose of satisfying customer requirements Process Customers Suppliers Resources Inputs Outputs
7
Process Types and Hierarchies
Individual processes Carried out by a single individual Make up 2. Vertical or Functional processes Contained within one functional unit or department Make up 3. Horizontal or Cross Functional processes Spans several functional units, departments or companies
8
Illustration: Process Types and Hierarchies
CEO Marketing Operations Accounting Buying a TV commercial Order Fulfilled Order Request Production planning Individual process Vertical process Horizontal process
9
Process Types and Hierarchies
Core cross-functional processes often have highest improvement potential Core processes – essential for meeting market place demand through a specific strategy Especially high improvement potential if a significant amount of non-manufacturing/service related activities Reasons Difficult to coordinate Have not kept up with improvements in manufacturing Difficult to detect waste and inefficiencies Often as little as 5% of the time considered adding customer value Customers more likely to abandon business because of “poor” service than “poor” products
10
Components of the Process Architecture
Inputs and Outputs Information structure Process Architecture Flow units Resources The network of activities and buffers
11
Components of the Process Architecture
Inputs and Outputs Establish interaction between the process and its environment Identify the process boundaries easy to identify the Input consumed from the environment in order to produce the desired Output Process inputs and outputs can be Tangible (Ex. raw material, cash, products, customers) Intangible (Ex. Information, time, energy, services) Flow units A flow unit is a transient entity or a job that proceeds through the network of activities and buffers and exits the process as a finished output Typically, the identity of a flow unit changes across the process Examples of common flow units: materials, orders, files, documents, customers, products, cash, transactions… Flow rate – The number of jobs flowing through the process per time unit
12
Components of the Process Architecture
The network of activities and buffers The work performed on a job moving through a process can be divided into an ordered sequence of activities The buffers represent storage or waiting points where the job waits before moving to the next activity (queues, waiting rooms, etc.) Different types of jobs different paths through the network Defining activities is crucial in process analysis Tradeoff between process and activity complexity Process Complexity Individual Activity Complexity
13
Components of the Process Architecture
Resources Tangible assets utilized to perform activities in a process Can be divided into: Capital assets – real estate, machinery, equipment, IT systems… Labor – people and their knowledge and skills Resources are utilized while inputs are consumed Information structure Specifies the information required for making decisions and performing activities in a process Limited information availability is a common cause for process inefficiencies Information enables coordination!
14
Workflow Management Systems
Management of administrative processes in the field of Information Systems is often referred to as workflow management Workflow management systems Control actions taken on documents moving through a business process Workflow management software/systems are used to control who does what to a specific document Using our comprehensive process definition Process = Workflow
15
The Essence of Business Process Design
“How to do things in an efficient and effective way” An efficient process which does not deliver customer value is useless A well designed process does the right things, right! More formally… BPD is concerned with configuring the process architecture to satisfy customer desires in an efficient way Customers can be both internal and external Internal customer requirements must be aligned with the desires of the external customers in the business strategy We make a clear distinction between process design and implementation
16
The Essence of Business Process Design
BPD often most valuable when considering complex cross functional processes Challenging coordination issues Process inefficiencies often related to handing off work from one station or person to the next – introduces delays and errors The functional organization and division of labor paradigm dates back to Adam Smith and the late 1700’s Division of labor rationale: by focusing on fewer tasks Workers’ skill level goes up work faster No time lost when workers switch between tasks Workers well positioned to help develop better techniques and tools Drawback: more complex coordination issues when More complex products and services Customers demand more variety
17
Incremental Improvement vs. Process Design
Subtle difference – both approaches concerned with how to do things better Complement each other Incremental process improvement: (continuous improvement) Change that brings a process closer to its normal operating standards Does not question the fundamental assumptions and rules that define the current process design Deductive approach Business Process Design Creative in its nature Questions existing assumptions and rules Requires new perspectives to generate innovative solutions with potential for breakthrough improvements Inductive approach
18
Incremental vs. Radical Design Improvement
Time Incremental Radical Theoretical Capability Statistical Process Control
19
Example 1 – Claims Handling in a Large Insurance Company
Pilot project – claims handling for replacement of automobile glass Springboard for later, more ambitious redesign efforts Set up procedure The CEO appoints an executive sponsor to lead the project Team members are handpicked by the CEO and the sponsor The team creates a flowchart of the existing process Under the existing process the client may have to wait 1-2 weeks before being able to replace the damaged auto glass Goal – A radical overhaul and improvement of the process to shorten the client waiting time
20
Flowchart of the existing claims process
Example 1 Flowchart of the existing claims process Client Local independent agent Approved glass vendor Claims processing center Request additional information Pay Notify agent File claim Give instructions Forward claim Request quote Provide quote
21
Explanation of existing claims process
Example 1 Explanation of existing claims process Client notifies a local agent that she wishes to file a claim. She is given a claims form and is told to obtain a cost estimate from a local glass vendor. When the claims form is completed the local agent verifies the information and forwards the claim to a regional processing center. The processing center logs the date and time of the claim’s arrival. The data is entered into a computer-based system (for record keeping only) by a clerk. The claim is then placed in a hard copy file and passed on to a claims representative. a) If the claims representative is satisfied with the claim it is passed along to several others in the processing chain and eventually a check is issued and sent to the client. b) If there are problems with the claim the representative mails it back to the client for necessary corrections. 5. When the client receives the check she can go to the local glass vendor and replace the glass.
22
New Design Recommended by the Team
Example 1 New Design Recommended by the Team Client Claims processing center Approved glass vendor Call in claim Schedule repair Pay Notify
23
Procedural changes to the new process
Example 1 Procedural changes to the new process The Claims representative is given final authority to approve the claim. Long term relationships with a limited number of glass vendors enables the insurance company to leverage its purchase power to pre-negotiate low prices. Clients no longer have to collect estimates. Vendors are certified for quality, price, reliability, etc. The Client now contacts the claims representative directly instead of going via a local agent.
24
Structural changes to the new process
Example 1 Structural changes to the new process A new 24 hour hotline enables the client to speak directly to a claims representative at the regional processing center. The claims representative gathers data over the phone, enters the data into the computer and resolves any issues on the spot. He tells the client to expect a phone call from a certain glass vendor to arrange the replacement. The claims information is immediately available for accounting via a LAN system and they can start processing the check and send it to the vendor.
25
Benefits with the new redesigned process
Example 1 Benefits with the new redesigned process The client can have the glass replaced within 24 hours As opposed to 10 days The client has less work to do Only one phone call, no need for a cost estimate Problems are handled immediately when the claim is filed Problems with lost or mishandled claims virtually disappear Fewer people are involved in the process lower op. costs Long term relationships with glass vendors Savings of 30-40% on paid claims due to special discounts Consolidated monthly payments lower handling costs More consistent and reliable service Claims representative feels ownership of the process Does a better job
26
Linking BPD to overall Business Performance
Detailed definition is company specific Generally, performance must be measured against the stated objectives Profit maximizing firms Non-profit organizations Overarching objective is usually to maximize long term shareholder value A common objective is survival and growth while satisfying customer needs Maximize revenues and minimize costs Must use resources efficiently while understanding customer needs Satisfying customer needs in an efficient way Well designed business processes
27
Linking BPD to Strategy
A unifying theme that helps align decisions made in an organization Guides a business towards its stated goals Two strategy levels Corporate strategy – Which industry should the business be in? Business strategy – How should we compete in a given industry? Intensified competition in all industries a prerequisite for success is to be highly competitive, i.e. to have an effective business strategy True also for many non-profit organizations that compete for funds
28
Linking BPD to Strategy
An effective business strategy is based on understanding the organization’s Internal environment – its strengths and weaknesses External environment – the opportunities and threats Links between BPD and the internal environment Weaknesses – often relate to poorly designed processes Strengths – often relate to well designed processes Links between BPD and the external environment Prerequisite for designing effective processes is to understand the external environment (suppliers, customers and competitors) and its opportunities and threats
29
Linking BPD to Strategy
Strategic fit Match between the strategic position the firm wants to occupy in the external market and the internal capabilities to get there Effective BPD is needed to achieve this fit Market driven strategy to achieve strategic fit Identify desired strategic position Design processes to support this position Flexibility, adaptability Time to market considerations Process driven strategy to achieve strategic fit Identify process capabilities offering a competitive advantage Leverage these capabilities to reach a desirable strategic position
30
Why are Inefficient Processes Designed?
They are usually not designed - They just emerge Circumstances and the process environment change and processes are incrementally adapted, but often without careful analysis of the overall effects Examples: see Laguna & Marklund Section 1.4
31
Activity Classification and BPD
A key issue in process design and analysis is classification of the process activities Crucial in identifying waste and inefficiencies in existing processes Two basic classification approaches: Activity Value-Adding Non-Value Adding Handoff Delay Rework Business Value Adding Control Policy compliance Value-Adding Activity Non-Value Adding Handoff Delay Rework Control Policy compliance
32
Activity Classification and BPD
Value adding activities Essential in order to meet customer expectations Activities the customer would be willing to pay for Involves doing the right things right Performing the right activities Doing them correctly, with high efficiency Business value adding activities Control activities Do not directly add customer value but are essential to conducting business Non-value adding activities Activities the customer is not willing to pay for
33
Activity Classification and BPD
Elimination of non-value adding activities is a key first step in redesigning business processes Often achieved through task or activity consolidation Task and activity consolidation reduces Hand-offs Need for control activities Process complexity
34
Process Management and Process Oriented Improvement Programs
Chapter 2 Business Process Modeling, Simulation and Design
35
Overview Process Management and the Power of Adopting a Process view
Six Sigma Definitions Cost and revenue rationale Framework Key success factors Business Process Reengineering What is it? Brief history What processes should be reengineered, and when? Suggested frameworks Evolutionary vs. Revolutionary Change
36
Process Management Principles of managing, controlling and improving processes Workflow oriented how jobs flow through an organization Important elements in managing processes Process design Continuous (incremental) improvement Control systems People management Change management Origins in the field of quality management Process control is a fundamental component Historically strong manufacturing focus Equally valuable in services
37
The Power of Adopting a Process View
Weaknesses of the functional org. and division of labor paradigm Focus on skills and resource utilization rather than work output Reward systems tailored for the functional unit not the overall firm Group behavior and cultures fostering an “us versus them” mentality Decentralization “firms within the firm” with their own agenda Strengths of a process view Creates focus on work output reduced risk for sub-optimization Leads to transparency of how contributions of individual workers fit into the “big picture” encourages involvement and empowerment Helps break down barriers between departments Creates a sense of loyalty towards the process to balance the loyalties within the functional units
38
Principles for Successful Process Management
Assign process ownership Perform feedback & control Develop & implement measures Establish control points Define the process Phase I: Initialization Phase II: Definition Phase III: Control Process authority, scope, interfaces and handoffs are determined Workflow documentation Baseline for process evaluation is defined Means and procedures for process monitoring, feedback and control are established Analyze boundaries & interfaces
39
Phase I: Initialization
Objective: Clarify the process scope Determine who will take responsibility for the process Process ownership Need someone in charge to make things happen Responsibilities of a process owner Accountability and authority for process operations and improvements Facilitate problem solving and make sure corrective action is taken Mediate between line managers with overlapping authorities Guidelines for assigning process ownership Manager with most resources or most work invested in the process Manager that is most affected if the process fails Process owner must have high enough position to see how the process fits into the “big picture”, needs clout to solve functional bickering
40
Phase I: Initialization
Analyzing Process Boundaries and Interfaces Process Boundary defines the process entry and exit points where inputs flow in and outputs flow out Provides a clear picture of the process scope Defines the external interfaces Internal interfaces Hand-off points within the process boundaries Most critical where the process crosses functional or organizational borders Most process inefficiencies are related to insufficient interface communication (= lack of coordination) Important to identify critical interfaces early on
41
The Customer-Producer-Supplier (CPS) model
Useful approach for resolving interface related problems Applying the CPS model to all critical interfaces adopt a view of the process as a chain of customers Coordination achieved by understanding internal & external customers Involves negotiation and agreement between all parts Producer PROCESS Customer Supplier Input Output Customer Requirements Producer Requirements Output Interface Input Interface
42
Phase II: Definition Define the process Objective:
Understand and document the process workflow Facilitate communication and understanding of process operations Define the process Documentation of work content in individual activities Usually in terms of verbal descriptions Operating procedures or Standard Operating Procedures (SOP) Documentation of process flows Usually a flowchart based method Combination of verbal and graphical description Common information gathering techniques Interviews with people working in the process (group or individual) Analytical observation Review of relevant documentation
43
Phase III: Control (I) Establish Control Points Objective:
Establish a system for controlling the process and providing feedback to the people involved Establish Control Points Control points are activities such as Inspection, verification, auditing, measuring, counting… Usually considered business value adding Without control points and a control system the only way of assessing process performance is customer feedback The process ends up in a reactive mode Poor quality is discovered too late Location of control points is determined by Criticality – impact on customer satisfaction Feasibility – physically and economically possible
44
Phase III: Control (II)
Develop and Implement Measurements Involves answering the questions What is to be measured and controlled (Ex. FedEx)? What is currently measured (available data)? Can a business case be made for a new measurement system? What is the appropriate sampling method, sampling size and frequency? Measurements should be meaningful, accurate and timely Statistical and graphical tools needed to turn data into information. Five measurement categories: Measures of… Conformance (to given specifications) Response time (lead-time, cycle time) Service levels (degree of availability) Repetition (frequency of recurring events such as rework) Cost (Quality, PAF, internal and external failure costs)
45
Phase III: Control (III)
Performing Feedback and Control Of critical importance for stabilizing and improving the process Objectives of control/corrective action are Regulation to maintain a certain performance level Improvement aiming at reducing variability or raising the average performance level Feedback is an important enabler for corrective action People in the process need to understand how their actions affect the overall process and its performance Feedback should be performed in a constructive – not punitive – manner Constructive feedback Makes people feel that they matter Encourages involvement and commitment
46
Six Sigma Quality Programs
Six Sigma is originally a company wide initiative at Motorola for breakthrough improvement in quality and productivity Launched in 1987 Rendered Motorola the Malcom Baldridge National Quality Award 1988 The ongoing success of Six Sigma programs has attracted a growing number of prestigious firms to adopt the approach Ex. Ford, GE, AMEX, Honeywell, Nokia, Phillips, Samsung, J.P. Morgan, Maytag, Dupont… Broad definition of Six Sigma programs “A company wide strategic initiative for process improvement in both manufacturing and service organizations with the clear objective of reducing costs and increasing revenues” Fierce focus on bottom line results
47
Technical Definition of Six Sigma
Reduce the variation of every individual process to render no more than 3.4 defects per million opportunities Assuming the process output is normally distributed with mean and standard deviation the distance between the target value and the closest specification limit is at least 6 and the process mean is allowed to drift at most 1.5 from the target Target Value (T) Upper Specification Limit (USL) Lower Specification Limit (LSL)
48
The Six Sigma Cost or Efficiency Rationale
Reducing costs by increasing process efficiency has an immediate effect on the bottom line To assure worker involvement Six Sigma strives to avoid layoffs Commitment Reduced Costs Increased Profits Improvement projects Cycle Time Yield Variation The Six Sigma Efficiency loop
49
The Six Sigma Cost or Efficiency Rationale
Oriented around the dimensions of variation, cycle time & yield Variation Can be divided into two main types Common cause or random variation Special cause or non-random variation Non-random variation Relatively few identifiable root causes First step in reducing the overall variation is to eliminate non-random variation by removing its root causes Random variation The result of many different causes Inherent in the process and can only be affected by changing the process design
50
The Six Sigma Cost or Efficiency Rationale
Variation (cont.) Important concepts in understanding the impact of variation Dispersion Predictability Centering Magnitude of variation in the measured process characteristics. Do the measured process characteristics belong to the same probability distribution over time? For a predictable process the dispersion refers to the width of the pdf. How well the process mean is aligned with the process target value.
51
The Six Sigma Cost or Efficiency Rationale
Variation (cont.) Ideally the process should be predictable, with low dispersion, and well centered Standard approach for reducing variability in Six Sigma programs Eliminate special cause variation to reduce overall dispersion and improve predictability Reduce dispersion of the predictable process Center the process to the specified target Six Sigma use traditional tools for quality and process control/analysis Basic statistical tools for data analysis The 7 QC tools
52
The Six Sigma Cost or Efficiency Rationale
Cycle time and Yield Cycle time (lead-time, response time) The time a job spends in the process Yield (productivity) Amount of output per unit of input or per unit time Improvement in cycle time and yield follow the same tactic as for variation Gain predictability, reduce dispersion and center to target The target is usually broadly defined as Minimize cycle time and Maximize yield Six Sigma principle Improvement in average cycle time and yield should not be made at the expense of increased variation
53
The Six Sigma Revenue or Effectiveness Rationale
Determinants of the company’s revenues Sales volume closely related to market share Sales prices Revenues contingent on how well the firm can satisfy the external customers’ desires An important Six Sigma Success factor is the focus on internal and external customer requirements in every single improvement project
54
The Six Sigma Cost & Revenue Rationale
Commitment Reduced Costs Increased Profits Improvement projects Cycle Time Yield Variation Customer satisfaction Increased Market Share & potentially higher prices Increased Revenues
55
The Six Sigma Framework
Centered around a disciplined and quantitatively oriented improvement methodology (DMAIC) Define, Measure, Analyze, Improve, Control Define Measure Analyze Improve Control Top Management Commitment Training Improvement Methodology Measurement System Stakeholder Involvement
56
Six Sigma Success Factors
The bottom line focus and big dollar impact Encourages and maintains top management commitment The emphasis on - and consistent use of - a unified and quantitative approach to process improvement The DMAIC methodology provides a common language so that experiences and successes can be shared through the organization Creates awareness that decisions should be based on factual data The emphasis on understanding & satisfying customer needs Creates focus on doing the right things right Anecdotal information is replaced by factual data The combination of the right projects, the right people and the right tools Careful selection of projects and people combined with hands on training in using statistical tools in real projects
57
Introduction to Reengineering
Business Process Reengineering (BPR) One of the buzzwords of the late 80’s and early 90’s “…achieves drastic improvements by completely redesigning core business processes” BPR has been the subject of numerous articles and books; classical examples are: “Reengineering Work: Don’t Automate, Obliterate”, Michael Hammer, Harvard Business Review, 1990 “The New Industrial Engineering”, Davenport and Short, Sloan Management Review, 1990
58
BPR Success Stories and Failures
Ford cuts payable headcount by 75% Mutual Benefit Life improves underwriting efficiency by 40% Xerox redesigns its order fulfillment process and improves service levels by 75-97% and cycle times by 70% with inventory savings of $500 million Detroit Edison reduces payment cycles for work orders by 80% Failures An estimated 50-70% of all reengineering projects have failed Those that succeed take a long time to implement and realize
59
Reasons for BPR Failures
Lack of support from senior management Poor understanding of the organization and the infrastructure Inability to deliver necessary technology Lack of guidance, motivation and focus Fixing a process instead of changing it Neglecting people’s values and beliefs Willingness to settle for marginal results Quitting too early Allowing existing corporate cultures and mgmt attitudes to prevent redesign Not assigning enough resources Working on too many projects at the same time Trying to change processes without making anyone unhappy Pulling back when people resist change Etc…
60
What does it take to succeed with BPR?
Hammer and Champy “The role of senior management is crucial.” Empirical research indicates… organizations which display understanding, commitment and strong executive leadership are more likely to succeed with process reengineering projects. Common themes in successful reengineering efforts Firms use BPR to grow business rather than retrench Firms emphasize serving customers & compete aggressively with quantity & quality of products & services Firms emphasize getting more customers, more work and more revenues instead of downsizing
61
Reengineering and its Relationships to Other Improvement Programs (I)
Reengineering - what is that? “The fundamental rethinking and radical redesign of business processes to achieve dramatic improvements in critical, contemporary measures of performance such as cost, quality, service and speed” (Hammer and Champy 1993) A number of similar definitions by other authors also exist Reengineering characteristics Focus on core competencies or value adding business processes The goal is to achieve dramatic improvement through rapid and radical redesign and implementation Projects that yield only marginal improvement and drag out over time are failures from a reengineering perspective
62
Reengineering and its Relationships to Other Improvement Programs (II)
Rightsizing Restructuring Automation TQM Reengineering Assumptions Staffing Reporting Technology Customer Fundamental questioned relationships applications needs Focus of Staffing, job Organization Systems Bottom-up Radical change responsibilities improvements changes Orientation Functional Functional Procedures Processes Processes Role of IT Often blamed Occasionally To speed up Incidental Key emphasized existing systems Improvement Usually Usually Incremental Incremental Dramatic and goals incremental incremental significant Frequency Usually one Usually Periodic Continuous Usually one time one time time
63
Relationship between Discontinuous (Radical) and Continuous Improvement
Time Incremental Radical Theoretical Capability Statistical Process Control
64
Brief History of BPR (I)
Most agree that Michael Hammer laid the foundation to the reengineering approach… …But many factors influenced the birth and hype around BPR The origins can be traced back to a number of successful projects undertaken by management consulting firms like McKinsey in the 80’s TQM had brought the notion of process improvement onto the management agenda The recession and globalization in late 1980’s and early 1990’s stimulated companies to seek new ways to improve business performance Programs often aimed at increasing flexibility and responsiveness Middle management under particular pressure
65
Brief History of BPR (II)
…But many factors influenced the birth and hype around BPR The Productivity Paradox (Stephen Roach) Despite powerful market and service innovations related to IT and increased computer power in the 1980’s there was little evidence that IT investments improved overall productivity Organizations were not able to utilize the capabilities of the new technology – Automating inefficient processes has limited impact on productivity Articles and books by Hammer, Davenport, Short, Champy etc. legitimized and defined the reengineering approach Early success stories were heavily published in the popular press Many consultants/vendors launched their own versions of BPR All types of change programs were labeled reengineering Gave BPR a bad name
66
When Should a Process be Reengineered? (I)
Three forces are driving companies towards redesign (The three C’s, Hammer & Champy, 1993) Customers are becoming increasingly more demanding Competition has intensified and is harder to predict Change in technology constant pressure to improve; design new products faster flexibility and ability to change fast are requirements for survival
67
When Should a Process be Reengineered? (II)
Useful questions to ask (Cross et al. (1994)) Are customers demanding more for less? Are your competitors providing more for less? Can you hand-carry a job through the process much faster than the normal cycle time (ex five times faster)? Have your incremental improvement efforts been stalled? Have technology investments been a disappointment? Are you planning to introduce radically new products/services or to serve new markets? Are you in danger of becoming unprofitable? Have cost-cutting programs failed to turn the ship around? Are operations being merged or consolidated? Are the core business processes fragmented?
68
What Should be Reengineered? (I)
Processes (not organizations) are reengineered Confusion arises because organizational units are well defined, processes are often not. Formal processes are prime candidates for reengineering Formal processes are guided by written policies; informal processes are not. Typically involve several departments and many employees. More likely rigid and therefore more likely to be based on invalid assumptions.
69
What Should be Reengineered? (II)
Screening criteria Dysfunction Which processes are in deepest trouble (most broken or inefficient)? 2. Importance Which processes have the greatest impact on the company’s customers? 3. Feasibility Which processes are currently most likely to be successfully reengineered?
70
Dysfunctional or Broken Processes
Symptoms and diseases of broken processes Symptom Disease 1 Extensive information Arbitrary fragmentation exchange, data redundancy of a natural process and re-keying 2 Inventory, buffers and System slack to cope with other assets uncertainty 3 High ratio of checking and Fragmentation control to value-adding 4 Rework and (re)iteration Inadequate feedback along chains 5 Complexity, exceptions Accretion onto a simple base and special cases
71
Importance Assessed by determining issues the customers feel strongly about and identifying which processes most influence these issues Customer Issues Product Cost On-time Delivery Product Features After-sales service Market Processes Product Design Order Processing Procurement CRM Company
72
Feasibility Determined by: Process Scope, Project Cost, Owner Commitment and the Strength of the Redesign Team Larger projects offer potentially higher payoffs but lesser likelihood of success Process Scope Project Cost Process Feasibility Owner/Corp. Commitment Team Strength
73
The Process Paradox “Get the right processes right”
The process paradox refers to the decline and failure of businesses that have achieved dramatic improvements through process reengineering To avoid getting caught in the process paradox companies must “Get the right processes right”
74
Suggested Framework for BPR (I)
In general, keywords for successful BPR are creativity and innovation… …but BPR projects also need structure and discipline, preferably achieved by following a well thought-out approach. BPR Framework due to Roberts (1994) Starts with a gap analysis and ends with a transition to continuous improvement. The gap analysis focuses on three questions: The way things should be The way things are How to reconcile the gap between 1 and 2
75
Robert’s Framework for BPR
Opportunity assessment Current capability analysis Process Design Risk and impact Transition plan Pilot test Infrastructure modifications Implementation and transition Tracking and performance Continuous improvement process
76
Suggested Framework for BPR (II)
BPR Framework due to Lowenthal (1994) Consists of 4 phases Preparing for change 3. Designing for change Planning for change 4. Evaluating change Phase 1 – Goals Building management understanding, awareness and support for change Preparing for a cultural shift and acquire employee “buy-in” Phase 2 – Assumption Organizations need to adopt to constantly changing marketplaces Phase 3 - Method To identify, assess, map and design A framework for translating process knowledge into leaps of change Phase 4 – Means Evaluate performance during a specified time frame
77
Lowenthal’s Framework for BPR
Preparing for change Planning for Designing Evaluating Phase I Phase II Phase III Phase IV
78
Suggested Framework for BPR (III)
BPR Framework due to Cross, Feather&Lynch (1994) Analysis In depth understanding of market and customer requirements Detailed understanding of how things are currently done Where are the strengths and weaknesses compared to the competition 2. Design Based on principles that fall into six categories Service Quality – relates to customer contacts Workflow – managing the flow of jobs Workspace – ergonomic factors and layout options Continuous improvement – self sustaining Workforce – people are integral to business processes Information technology Implementation Transforming the design into day to day operations
79
Cross et al’s Framework for BPR
Customer Requirement analysis Design specifications High-level design Detailed design Pilot new design Transform the business Baseline analysis Current process review Design options Model/validate new design Build in CI feedback principles Analysis Phase Design Phase Implementation Phase
80
Revolutionary vs. Evolutionary Change
The reengineering movement advocates radical redesign and rapid revolutionary implementation and change A revolutionary change tactic Turns the whole organization on its head Has potential to achieve order of magnitude improvements Is very costly Has a high risk of failure To reduce risks and costs of implementation many companies end up with a strategy of radical redesign and evolutionary implementation tactic Implementing the feasible plans given current restrictions Implemented process is usually a compromise between the original process and the “ideal” blueprinted process design
81
Revolutionary vs. Evolutionary Change
Elements of evolutionary and revolutionary change theories Element Evolutionary change model Revolutionary change model Leadership Insiders Outsiders Outside resources Few, if any, consultants Consultant led initiative Physical separation No, part time team members Yes, “off-campus site” Crisis None Poor performance Milestones Flexible Firm Reward system Unchanged New IT/process change Process first Simultaneous process and IT change The critical element in choosing between a revolutionary and evolutionary approach is time If the firm is in a reactive mode responding to a crisis a revolutionary approach may be the only option If in a proactive mode an evolutionary tactic might work
82
The Evolutionary Change Model (I)
Basic principle People directly affected by or involved in a change process must take active part in the design and implementation of that change Real change is achieved through incremental improvement over time Change should come from within the current organization Should be carried out by current employees and leadership Should be adapted to existing resources and capabilities flexible milestones Should be based on open and broad communication New processes and procedures are implemented before introducing new IT systems
83
The Evolutionary Change Model (II)
Advantages of an evolutionary change tactic compared to a revolutionary approach Less disruptive and risky Increases the organization's ability to change Disadvantages Takes a long time to see results Does not offer the same potential for order of magnitude improvements Vision must be kept alive and adjusted over time as external market conditions change
84
The Revolutionary Change Model (I)
Based on the punctuated equilibrium paradigm Radical change occurring at certain instances Long periods of incremental change in between Revolutionary change Happens quickly Alters the very foundation of the business and its culture Brings disorder, uncertainty, and identity crises Needs to be top driven Requires external resources and new perspectives Involves tough decisions, cost cutting and conflict resolution The change team is small and isolated from the rest of the organization Avoid undue influence from current operations Communication with people in the process is on a “need to know” basis
85
The Revolutionary Change Model (II)
Advantages with a revolutionary implementation approach Drastic results can be achieved quickly If successful, the ideal “blueprinted” design is put in place Disadvantages with a revolutionary change tactic Very strenuous for the organization High probability for failure Diverts top management attention from the external marketplace Goes against core values of many organizations Empowerment Bottom-up involvement Innovation Secrecy creates uncertainty about the future roles of individual employees resistance to change
86
A Simulation Based Framework for Business Process Design Projects
Chapter 3 Business Process Modeling, Simulation and Design
87
Overview The Overall Framework
Step 1: Case for Action and Vision Statement Step 2: Process Identification and Selection Step 3: Obtain Management Commitment Step 4: Evaluate Design Enablers Step 5: Acquire Process Understanding Step 6: Creative Process Design Benchmarking Design Principles Step 7: Process Modeling and Simulation Step 8: Implementation of the New Process Design
88
A Simulation Based Framework for BPD Projects
1. Case for action and vision 2. Process Identification and selection 3. Obtaining Management commitment 4. Evaluation of Design Enablers 5. Acquiring Process understanding 6. Creative Process Design 7. Process Modeling and Simulation 8. Implementation of the New Process Design
89
Case for Action and Vision Statements (I)
A clear message about the need for change and where the change is going to take us is necessary for successfully selling the redesign concept to the company’s employees Case for Action Here is where we are as a company and this is why we cannot stay here Five major elements build an effective argumentation Business context – what is important and what is changing Business problems – source of the company’s concern Marketplace demand – performance standards & demands to meet Diagnostics – why are we unable to meet the posed demands Cost of inaction – consequences of not changing
90
Case for Action and Vision Statements (II)
Example: Case for action in a pharmaceutical company We are disappointed by the length of time we require to develop and register drugs in the United States and in major international markets. Our leading competitors achieve significantly shorter development cycles because they have established larger-scale, high-flexible, globally integrated R&D organizations that operate with a uniform set of work practices and information systems. The competitive trend goes against our family of smaller, independent R&D organizations, which are housed in several decentralized operating companies around the world. We have strong competitive and economic incentives to move as quickly as possible toward a globally integrated model of operation. Each week we save in the development and registration process extends the commercial life of our patent protection and represents, at minimum, an additional $1 million in annual pretax profit — for each drug in our portfolio.
91
Case for Action and Vision Statements (III)
This is what we as a company need to become Should include both quantitative and qualitative statements Need not be excessively long but should not be simplistic
92
Case for Action and Vision Statements (IV)
Example: Vision Statement in a pharmaceutical company We are a worldwide leader in drug development. We have shortened drug development and registration by an average of six months. We are acknowledged leaders in the quality of registration submissions. We have maximized the profit potential of our development portfolio. We have created, across our operating companies, a worldwide R&D organization with management structures and systems that let us mobilize our collective development resources responsibly and flexibly. We have established uniform and more disciplined drug development, planning, decision-making, and operational processes across all sites. We employ innovative technology-based tools to support our work and management practices at all levels and between all R&D sites. We have developed and implemented common information technology architecture worldwide.
93
Process Identification and Selection
Process selection is critical for the success of a design project Core processes have the highest impact on overall performance but are also more costly and risky to change The implementation tactic cannot be ignored, even due to budget constraints Useful criteria for prioritization of projects are: Dysfunction Importance Feasibility Other relevant screening issues/questions are: What are the project’s scope and costs involved? Can a strong and effective team be formed? Is it likely to obtain strong management commitment? Can other programs (e.g. continuous improvement) be used instead? Is the process obsolete or the technology outdated?
94
Obtaining Management Commitment
Top management must set the stage both for the design project and the subsequent implementation Without top management support the improvement effort is bound to fail The more profound and strategic the change is the more crucial the top management support becomes Commitment assumes understanding and cannot be achieved without education People are more likely to be fearful and resisting change if there is a lack of direction and they do not understand the implications of the change Occurrence of “resisting change” issues is particularly prevalent in rapid revolutionary change scenarios
95
Evaluation of Design Enablers
New (information) technology is an essential design enabler… …but could also reinforce old ways of thinking Automation redesign Do not look for problems first and then the technology to fix them Evaluating new technology needs inductive thinking New technology should not be evaluated within the structure of the existing process New technology enables us to break old rules and compromises To avoid the automation trap the question to ask is: How can new technology enable us to do new things or to do things in new ways?
96
Technology as a Mechanism to Break Rules and Compromises
Old Rule New Technology New Rule Information can appear Shared databases in only one place at a time. simultaneously in as many places as needed. Only experts can perform Expert systems A generalist may be able to do complex work. the work of an expert. Businesses must choose Telecommunication Businesses can simultaneously Between centralization networks reap the benefits of centralization and decentralization. Managers make all decisions. (databases, modeling tools) everyone’s job. Decision support tools - Decision making is part of Field personnel need offices Wireless data Field personnel can send and Where to receive, store, retrieve portable computers. receive information wherever and transmit information. communication and they are. The best contact with a potential Interactive videodisk buyer is personal contact. and web pages. buyer is effective contact. People must find where Automatic identification Things tell you where they are. tracking technology. Plans get revised instantaneously. things are. Plans get revised periodically. High performance computers. Plans get revised instantaneously.
97
Acquiring Process Understanding
Subtle difference between redesigning an existing process and designing a new currently non-existing process In both cases we need to understand the purpose of the process and what the customers desire from it If the process exists, we need to understand what it is currently doing and why it is unsatisfactory Business Process Benchmarking may be a useful tool To gain process understanding To inspire creative new designs
98
Understanding the Existing Process
Questions the design team needs to answer What is the existing process doing? How well does it perform? What are the critical issues that impact the process performance? The redesign team must understand the process but should not overanalyze it in order to avoid “analysis paralysis” Becoming so familiar with the process it is impossible to think of new ways of doing it Essential activities for building process understanding Configure the redesign team Build a high level process map Test the initial scope and scale Identify the process owner
99
Activities for Building Process Understanding (I)
Configure the redesign team A mix of business insiders (managers and workers directly involved in the current process) and business outsiders (consultants and employees not involved in the process) Build a high level process map Neither a low level flow chart nor an organizational chart Shows interactions between sub-processes, not the flow of data Focuses on customers and business outcomes Objectives 1. Build common understanding 4. Use a cross functional vocabulary 2. Highlight critical sub-processes 5. Test initial scope and scale 3. Identify key interfaces 6. Pinpoint redundancies and waste
100
High Level Process Map for a Telecom Company
Mass Markets Service Delivery Service Assurance Local Network Operations Customers Capacity Provisioning Customer Transactions and Billing Carrier Service Delivery Markets & Planning
101
Activities for Building Process Understanding (II)
3. Test the initial scope and scale Self examination Environmental scanning/benchmarking Customer visits 4. Identify the process owner The person that will take responsibility and be accountable for the performance of the new process
102
Understanding the Customer
The customer end is the best place to start understanding a business process What are the customers’ real requirements? What do they say they need and what do they really need? What problems do they have? What do they do with the process output? The ultimate goal with a business process is to satisfy the customers’ real needs in an efficient way!
103
Creative Process Design (I)
Designing new processes is more of an art than a science Cannot be achieved through a formalized method Most existing processes were not designed; they just emerged as new parts were added iteratively to satisfy immediate needs The end result of any design is very much dependent on the order in which information becomes available Inefficient processes are created when iterative design methods are applied
104
Illustration Process Evolution (I)
Two pieces of plastic are given to you with instructions to arrange them in an easily described shape
105
Illustration Process Evolution (II)
Then a third piece is added still the objective is to build a simple shape
106
Illustration Process Evolution (III)
Two more pieces are added, but very few people are able to incorporate these and still obtain a simple shape ?
107
Illustration Process Evolution (IV)
Considering the pieces independently of the sequence by which they appear leads to a much better solution!
108
Benchmarking Comparing the firm’s/process’s activities and performance with what others are doing In the same company, in the same industry or across industries Every benchmarking relationship involves two parties The initiator firm – who initiates contact and observes (the pupil) The target firm (or benchmark) – who is being observed (the master) Fruitful benchmarking relationships are usually characterized by reciprocity Two basic benchmarking purposes To assess the firm’s/process’s performance relative to the competition identify performance gaps and goals To stimulate creativity and inspire innovative ideas for how to do things better, i.e. improve process designs & process performance For BPD projects both purposes are relevant
109
Business Process Benchmarking (I)
Focus on how things are done Typically the most involved type of benchmarking The underlying idea is to learn and be inspired by the best The best in a certain industry (best-in-class benchmark) The best across industries (best-of-the-best benchmark) Generally, the further away from the firm’s own industry that the design team goes Higher potential for getting breakthrough design ideas More difficult to identify and translate similarities between processes After choosing a target firm a good starting point for a business process benchmarking effort is the 5w2h framework (Robinson 1991) Can also be used to understand an existing process to be redesigned
110
Business Process Benchmarking (II)
The 5w2h framework Classification 5w2h questions Description People Who? Who is performing the activity? Why is this person doing it? Could/Should someone else perform the activity? Subject matter What? What is being done in this activity? Can the activity in question be eliminated? Sequence When? When is the best time to perform this activity? Does it have to be done at a certain time? Location Where? Where is this activity carried out? Does it have to be done at this location? Purpose Why? Why is this activity needed? Clarify its purpose. Method How? How is the activity carried out? Is this the best way or are there alternatives? Cost How much? How much does it currently cost? What would be the tentative cost after improvement?
111
General people-oriented and conceptual process design principles
1. Organize work around outcomes, not tasks 2. Let those who use the process output perform the process 3. Merge information processing and data gathering activities 5. Put the decision point where the work is performed and build control into the process 4. Capture the information once – at the source 8. Design the process for the dominant flow not the exceptions 6. Treat geographically dispersed resources as though they were centralized 7. Link parallel activities instead of just integrating their output 9. Look for ways to mistake-proof the process 10. Examine process interactions to avoid sub-optimization Themes: Horizontal and vertical integration of work, hand-off elimination, improved quality and task coordination Coordination of activities, simplification of flows, elimination of waste and rework
112
Ten Conceptual Design Principles (I)
Organize work around outcomes not tasks Focus on horizontal integration of activities Eliminates unnecessary handoff and control steps Process complexity is reduced while activity complexity grows This integration approach often referred to as case management Let those who use the process perform the process Work should be carried out where it makes most sense to do it Risk of coordination inefficiencies due to excessive delegation decreases Merge information processing and data gathering activities The people collecting the data should also process it into information Reduces the risk of errors and incorrect information 4. Capture information once – at the source Reduces costly reentry and frequency of erroneous data Speeds up the process, increases the quality of information and reduces costs
113
Ten Conceptual Design Principles (II)
5. Put the decision point where the work is performed and build control into the process Case management compresses processes horizontally and employee empowerment compresses them vertically Workers are taking over previous management responsibilities Treat geographically dispersed resources as though they were centralized IT breaks spatial compromises through virtual co-location Geographically disbursed resources should not constrain the design team to only consider decentralized approaches Link/coordinate parallel activities instead of just integrating their results If parallel activities are operated independently operational errors are not detected until the outcomes are integrated Reduces the amount of rework
114
Ten Conceptual Design Principles (III)
8. Design the process for the dominant flow not for the exceptions Reduces the risk of fragmentation and overly complex processes with inherent coordination problems 9. Look for ways to mistake-proof (or fail-safe) the process Design so that certain critical errors cannot occur Mistake-proofing = Poke Yoke Examining interactions to avoid sub-optimization By neglecting interactions, isolated improvements to sub-processes will lead to sub-optimal solutions Known in systems theory as “disjointed incrementalism”
115
Seven Workflow Oriented Design Principles
Stems from the field of industrial engineering Successfully used for designing manufacturing systems for decades Traditional, technically oriented workflow design principles Focus: Efficient process flows, managing resource capacity, throughput and cycle times Establish product orientation in the process Eliminate Buffers Establish one at a time processing Balance the flow to the bottleneck Minimize sequential processing and hand-offs Schedule work based on its critical characteristics Minimize multiple paths due to specialized operations for exception handling
116
Process Modeling and Simulation (I)
Conceptual process designs need to be tested before they are implemented in full scale Pilot projects or process modeling techniques Business processes are often too complex and dynamic to be analyzed only with simple tools like flowcharts and spreadsheets Discrete event simulation is a powerful and realistic tool to complement the more simplistic methods Allows exploration of the redesign effects without costly interruptions of current operations Helps reduce the risks inherent in any design/change project Compared to pilot projects simulation is faster and cheaper Simulation not good for capturing soft people issues and attitudes Simulation and pilots complement each other
117
Process Modeling and Simulation (II)
A discrete event simulation model mimics the real world but in compressed time Focus only on events when the state of the system changes and skips the time between these events Basic steps in evaluating a process design through discrete event simulation Building the simulation model Running the simulation Analyzing performance measures Evaluation of alternative scenarios
118
Process Modeling and Simulation (III)
Advantages with discrete event simulation Promotes creativity by enabling easy testing of ideas Captures system dynamics but avoids disturbances of current process Can capture interactions between sub-processes Mitigates the risk of sub-optimization Graphical reporting features promotes better process understanding and facilitates communication The quantitative nature brings a sense of objectivity into the picture
119
Implementation of the Process Design (I)
Detailed implementation issues beyond the scope of the design project High level implementation issues need to be considered when selecting a process to design No point in designing a process which cannot be implemented Crucial high level implementation issues Time Cost Improvement potential Likelihood of success
120
Implementation of the Process Design (II)
Conceptually an implementation strategy can be characterized as revolutionary, evolutionary or on a continuum in between A rapid revolutionary approach tends to require more external resources Regardless of the implementation tactic important factors for a successful implementation are Strong leadership Buy-in from line managers and employees Training of the workforce
121
Final Notes Important to reflect on what can be learned from a given design and/or implementation project What worked, what didn’t and why? What were the main challenges? What design ideas didn’t work out in practice and why? The process of designing and implementing new process designs also needs improvement Sharing experiences and collecting feedback is key to any improvement effort
122
Basic Tools for Business Process Design
Chapter 4 Business Process Modeling, Simulation and Design
123
Overview Introduction Basic Tools for BPD Graphical tools
General Process Charts Process Activity Charts Process Flow Diagrams Flow Charts Service System Mapping Workflow Design Principles and Tools Establish product orientation in the process Eliminate buffers One-at-a-time processing Balancing bottleneck flows Minimize sequential processing and handoffs Scheduling based on job characteristics Minimize multiple paths
124
Basic Tools for Process Design
Deterministic tools and modeling approaches to help designers analyze processes and check proposed designs for Feasibility Completeness Efficiency Quantitative tools require data regarding important process characteristics Steps required to complete the process Processing and activity times are key Tagging is an important technique for gathering process data Follow a job through the process Data is collected on a document (a tag) accompanying the job Particularly useful for gathering data on processing and activity times
125
General Process Charts
Summarizes the current process the redesigned process and the expected improvements Characterizes the process by The number of activities per category The amount of time spent in each activity category The percentage of the total processing time spent on each category Clearly indicates Major problems with the existing process How the redesigned process remedies these problems Problems measured in terms of the time and the percentage of time spent on non-value adding activities
126
Illustration of a General Process Chart
Activities Current Process Redesigned Process Difference No. Time % No. Time % No. Time Operations Inspections Transport Storage Delays Total
127
Process Activity Charts
Complements the general process chart Provides details regarding the sequence of activities Disadvantages Only considers average activity times If the process includes several variants with different paths (i.e. multiple paths through the process) each variant needs its own activity chart Cannot depict parallel activities
128
Illustration of a Process Activity Chart
Symbols Process Activity Chart Process: __________________________ Developed by: ____________________ Page: ____ of _____ Date: __________ Current Process Proposed Process Description Time Value code (V/N/C) Symbol No. For each activity, fill in the required information. Also, connect the symbols to show the flow through the process. The value code indicates whether the activity adds value (V), does not add value (N), or controls (C). Adoption 1 6 9/14/99 X Find where to go N Walk through V What’s next? N Operation Inspection Storage Delay Transportation of a physical item
129
Process Flow Diagrams (I)
Provide a picture of the spatial relationships between activities Typical application is for production floor layout problems. The diagram is used for measuring process performance in units of time and distance Including both horizontal and vertical movements. Assumes that moving items requires a time proportional to the distance. Can be used in conjunction with Process Activity Charts By labeling areas in the process flow diagram and by adding a column to the activity chart, indicating for each activity which area it belongs to. Alternatively, the flow diagram includes the activity numbers in the activity chart.
130
Illustration of a Process Flow Diagram
Before Redesign After Redesign C B A F E Incoming request Finished request D C B A F E D Finished request Incoming request
131
Process Flow Diagrams (II)
Analysis geared towards reducing excessive and unnecessary transportation and movements of items/jobs Long distances Crisscrossing paths Repeated movements between the same activities Other illogical flows Can be used as a basis for computing Load Distance (LD) scores Useful for quantitatively comparing alternative designs/layouts with regards to flow rates and distances
132
Load Distance Analysis
LD(i,j) = LD score between work centers i and j The LD score measures the attraction between two work centers (activities) The larger the traffic volume the higher the score and the higher the incentive to keep the work centers together The goal is to find a design that minimizes the total LD score (the sum of individual scores between work centers) The Load Matrix summarizes the load (flow rate = # of jobs) that needs to be shipped between each pair of work centers LD(i,j) = Load(i,j)*Distance(i,j)
133
A Sample Load Matrix A B C D E F A 20 20 80 B 10 75 C 15 90 D 70
See also example exercise on LD Analysis
134
Flow Charts One of the fundamental graphical tools for process analysis and design Typically depicts activities sequentially from left to right Can help to identify, loops, multiple alternative paths, decision points etc. Symbols often used in flow charting Operation Transportation of a physical item Storage Inspection Transportation of information Delay
135
Illustration of a Sample Flow Chart
Operator takes phone order. Orders wait to be picked up. Supervisor inspects orders. Order is fulfilled. Order waits for sales rep. Is order complete? Yes No Orders are moved to supervisor’s in-box. Orders wait for supervisor.
136
Activity Times and Path Frequencies (I)
Flowcharts may be used to estimate the total average process time from the estimated activity times Assumes that the standard processing time is known (estimated) Assumes that the standard setup time is known (estimated) The standard times assumes 100% worker efficiency. If the worker is less efficient the times must be adjusted as above. Average activity time = (Unit processing time)*(batch size)+setup time Efficiency
137
Activity Times and Path Frequencies (II)
Example: Inspection activity Inspection of one unit takes 3 minutes Each inspection batch includes 10 units It takes 15 minutes to prepare for the inspection of a batch The inspector is new on the job and it currently takes 25% longer time to inspect a batch than when she is fully trained What is the estimated activity time for inspecting a batch? The average activity time = ((3*10) + 15)/0.75 = 60 minutes
138
Service System Mapping (I)
An extension of traditional flowcharting Illustrates how effectively a business process satisfies customers across all encounters Documents the role played by the customer in the service delivery process A combination of service blue printing and traditional flowcharting Goals with SSM Build consistent perceptions of customer’s experience with core processes Identify all points of contact between the process and its customers Provide a basis for developing an economic business model Identify opportunities within the process Provide a design framework Aid in pinpointing control points and strategic performance measures
139
Service System Mapping
An extension of traditional flowcharting Illustrates how effectively a business process satisfies customers across all encounters Documents the role played by the customer in the service delivery process SSM Horizontal Bands The purpose is to organize activities according to the people or “players in the process. – Who does what? A SSM typically consists of 5 bands Customer band – end user Frontline or distribution channel band Back-room activity band Centralized support or information systems band Vendor or supplier band SSM Process Segments A process segment or sub process is a set of activities that produces a well defined output given some input
140
Sample SSM for an Order Fulfillment Process
Receiving Filling Shipping Billing Customer Band Front line Band Back-room Band IS Band Supplier Band
141
Additional Diagramming Tools
Software products for flowcharting and diagramming… Micrografx PaceStar RFFlow Visual Thought SmartDraw TeamFlow Tension Software Visio 2000 A specialized approach for “enterprise modeling and analysis” is the so called IDEF methodology A family of structured methods (functional, information and data modeling) Based on an established graphical language SADT (Structured Analysis and Design Technique) Used by many consulting firms not least in design of information systems
142
Workflow Design Principles and Tools (I)
Organizing Activities Two basic ways of organizing activities By process (Process Orientation) By product (Product Orientation) Process orientation (functional layout) groups activities or workstations according to function Most common when the same activity is used for producing different products or services or when serving many different customers Utilization of equipment and personnel tends to be high
143
Workflow Design Principles and Tools (II)
Product orientation groups all necessary activities to complete a finished product into an integrated sequence of work nodes or work stations A typical example is an assembly/production line for making a particular car model Activities are organized around the route (needs) of a particular product or service Advantages with product orientation include Faster processing rate Lower WIP Less unproductive time due to setups Less transportation time Less handoffs A capital intensive way of organizing activities
144
Product v.s. Process Orientation
Product Orientation 1 2 4 5 3 Customer B Customer A (a) 1 2 4 5 3 Customer B Customer A (b)
145
A Hybrid Orientation To justify a product orientation from a resource utilization perspective the product/service volumes must be quite high. A popular hybrid between product and process orientation in manufacturing is known as Group Technology (or product clustering) Groups products with similar characteristics into families and organizes activities around these families instead of around the individual products “Product Family” orientation The equivalent in business processes would be to group jobs with similar characteristics into families. The hybrid orientation simplifies customer routings, reduces process time and can be justified even if the volumes of individual products/services are not that large
146
Illustration of a Hybrid Orientation
1 2 4 5 3 Customer C Cell AC Cell BD Customer A Customer B Customer D
147
Workflow Design Principles and Tools (III)
Buffer Elimination Buffers are put in place to protect against variability in demand, processing times, etc. Jobs stacked up at different parts of the process, waiting to be processed. WIP = Work In Process inventories. All jobs currently in the process, i.e. in queues/buffers, under transportation or under processing. Buffers tend to cause logistical and communication problems due to slower information feedback. Implies the need for advanced tracking systems to identify what job is in which buffer. Product orientation implies less WIP but needs to be well balanced in order to minimize buffers.
148
Workflow Design Principles and Tools (IV)
One-at-a-time Processing Reduction of the batch size to the size of one unit By reducing batch sizes (and setup times) the throughput time and WIP can be minimized Two types of batches Process batch All jobs being processed before the resource needs to be changed to process jobs of a different kind Transfer batch Number of items/jobs transported together to the next resource for processing Usually the process batch By reducing the transfer batch total processing time and WIP are also reduced
149
Example – Effect of Reducing Batch Sizes
1 2 3 Process Batch Transfer 100 50 250 Three activities in sequence 1, 2 & 3 Processing times: 1 h/job in 1&3 and 0.5 h/job in activity 2 Consider the total throughput time for a batch of 100 units when the transfer batch size is: A) 100 B) 20 100 20 130 1 2 3 Process Batch Transfer
150
Workflow Design Principles and Tools (V)
Balancing bottleneck flows Linked to the OM principle known as Theory of Constraints (TOC) popularized by Eliyahu Goldrat in his book The Goal Balance flow not capacity! Keep bottlenecks fed! Historically manufacturers had tried to balance capacity across processes to match market demands Making all activity capacities the same makes sense only if processing times are constant or display marginal variability Variation in processing times causes inventory build up and idleness at different parts of the process Only two ways of handling variation Increase WIP to smooth variation Differentiate/balance capacity according to the job flows
151
Workflow Design Principles and Tools (VI)
Line Balancing A useful approach when processing times are fairly constant Should not be used when processing times display high variability The goal is to balance the capacity of the different workstations constituting the production line (the process) Procedure Specify sequential (precedence) relationships among the activities using a precedence diagram Use market demand to determine the line’s desired cycle time per work station (C) C = Process time per day Market demand per day (in # of jobs)
152
Workflow Design Principles and Tools (VII)
Line Balancing Procedure (continued) Determine the theoretical minimum # of workstations (TM) Select a primary rule to assign activities to workstations and a secondary rule to break ties Assign activities one at a time to workstation 1 as long as the sum of activity times C. Repeat this for workstations 2,3, … Must satisfy the activities’ precedence relationships Evaluate the line efficiency = Total process time/(C*#stations) 7. Rebalance using a different priority rule in case the efficiency is unsatisfactory TM = Sum of activity times C
153
Workflow Design Principles and Tools (VIII)
Potential Line Balancing Complications Market demand may require a work station cycle time shorter than the longest activity time Need to change the process in some way! Approaches: Split the activity Use parallel workstations Train the workers or upgrade machinery for faster processing time Work overtime Redesign the entire process
154
Workflow Design Principles and Tools (IX)
Minimize Sequential Processing and Handoffs Sequential processing implies longer process throughput time Operations are dependent constrained by the slowest activity No one person is responsible for the entire service encounter Illustrative example (see figure on next slide) A process with 4 activities, throughput time 30 minutes and processing times 10, 7, 8 & 5 minutes in the 4 activities Sequential set up – each individual performs a different activity The process output is 60/10=6 jobs per hour The efficiency = ( )/(10*4) = 75% Parallel set up – each individual performs all 4 activities The process output is now 4*(60/30) = 8 jobs per hour The efficiency = 30/30 = 100%
155
Illustrative Example – Sequential v.s. Parallel Processing
Sequential processing 10 min 7 8 5 6 jobs/hour Parallel processing 30 min 8 jobs/hour
156
Workflow Design Principles and Tools (X)
Scheduling based on job characteristics (I) Scheduling involves sequencing the order at which a number of different jobs are to pass through a workstation or process with limited capacity Becomes more important as the diversity of jobs increases Characteristics that are commonly used as a basis for scheduling Arrival time Estimated processing time Due date Importance
157
Scheduling Based on Job Characteristics (II)
Finding the “right”objective function and the best scheduling characteristic to satisfy this objective is tricky Three common overall objectives Maximize process output over a given time period Satisfying customer desires for quality and promptness Minimizing current out-of-pocket costs Common surrogate objectives that are easier to quantify Minimize the makespan (the throughput time for a defined set of jobs) Minimize total (or average weighted ) tardiness (the time by which the completion time surpasses the due date) Minimize the maximum tardiness Minimize the number of tardy jobs The weighted tardiness is obtained as the product between the tardiness value and the importance weight of the job in question
158
Scheduling Based on Job Characteristics (III)
Commonly used priority rules are First-In-First-Out (FIFO) – scheduling according to arrival times Earliest-Due-Date first (EDD) Shortest Processing Time first (SPT) (See example illustrating the application and effect of the different rules) Observations for a single server situation EDD render the optimal solution to the problem of minimizing the maximum tardiness SPT render the schedule that minimizes the average throughput time per job for a given set of jobs
159
Scheduling Based on Job Characteristics (IV)
Moore’s Algorithm A method for minimizing the number of tardy jobs, when all jobs are considered equally important 1. Order the jobs according to the EDD rule. Stop if no jobs are tardy – the optimal solution is found! Go to step 6. 3. Find the first tardy job in the sequence. 4. Assuming that this tardy job is the kth in the sequence. Find and remove job j (j=1, 2, 3, …, k) with the longest processing time. 5. Revise the completion times and return to step 2 6. Insert the removed jobs at the end of the sequence in the order they were removed See example illustrating the application of Moore’s algorithm
160
Workflow Design Principles and Tools (XI)
Minimize the number of multiple paths through the process Reduces complexity and confusion Simplifies resource management and scheduling Fewer jobs are routed the wrong path and need to be rerouted or reworked One way of reducing the number of paths without compromising the efficiency and customization ability is to use case teams, i.e., horizontal compression of work flow
161
Managing Process Flows
Chapter 5 Business Process Modeling, Simulation and Design
162
Overview Processes and Flows – Important Concepts Cycle Time Analysis
Throughput WIP Cycle Time Little’s Formula Cycle Time Analysis Capacity Analysis Managing Cycle Time and Capacity Cycle time reduction Increasing Process Capacity Theory of Constraints
163
Processes and Flows – Concepts
A process = A set of activities that transforms inputs to outputs Two main methods for processing jobs Discrete – Identifiable products or services Examples: Cars, cell phones, clothes etc. Continuous – Products and services not in identifiable distinct units Examples: Gasoline, electricity, paper etc. Three main types of flow structures Divergent – Several outputs derived from one input Example: Dairy and oil products Convergent – Several inputs put together to one output Example: Car manufacturing, general assembly lines Linear – One input gives one output Example: Hospital treatment
164
Process Throughput Inflow and Outflow rates typically vary over time
IN(t) = Arrival/Inflow rate of jobs at time t OUT(t) = Departure/Outflow rate of finished jobs at time t IN = Average inflow rate over time OUT = Average outflow rate over time A stable system must have IN=OUT= = the process flow rate = process throughput
165
Process Inflow and Outflow vary over time
166
Work-In-Process All jobs that have entered the process but not yet left it A long lasting trend in manufacturing has been to lower WIP by reducing batch sizes The JIT philosophy Forces reduction in set up times and set up costs WIP = Average work in process over time WIP(t) = Work in process at time t WIP(t) increases when IN(t)>OUT(t) WIP(t) decreases when IN(t)<OUT(t)
167
The WIP Level Varies With Process Inflow and Outflow
168
Process Cycle Time The difference between a job’s departure time and its arrival time = cycle time One of the most important attributes of a process Also referred to as throughput time The cycle time includes both value adding and non-value adding activity times Processing time Inspection time Transportation time Storage time Waiting time Cycle time is a powerful tool for identifying process improvement potential
169
Little’s Formula (Due to J.D.C. Little (1961))
States a fundamental and very general relationship between the average: WIP, Throughput (= ) and Cycle time (CT) The cycle time refers to the time the job spends in the system or process Implications, everything else equal Shorter cycle time lower WIP If increases to keep WIP at current levels CT must be reduced A related measure is (inventory) turnover ratio Indicates how often the WIP is entirely replaced by a new set of jobs Little’s Formula: WIP = ·CT Turnover ratio = 1/CT
170
Cycle Time Analysis The task of calculating the average cycle time for an entire process or process segment Assumes that the average activity times for all involved activities are available In the simplest case a process consists of a sequence of activities on a single path The average cycle time is just the sum of the average activity times involved … but in general we must be able to account for Rework Multiple paths Parallel activities
171
Rework Many processes include control or inspection points where if the job does not conform it will be sent back for rework The rework will directly affect the average cycle time! Definitions T = sum of activity times in the rework loop r = percentage of jobs requiring rework (rejection rate) Assuming a job is never reworked more than once Assuming a reworked job is no different than a regular job CT = (1+r)T CT = T/(1-r)
172
Example – Rework effects on the average cycle time
Consider a process consisting of Three activities, A, B & C taking on average 10 min. each One inspection activity (I) taking 4 minutes to complete. X% of the jobs are rejected at inspection and sent for rework What is the average cycle time? If no jobs are rejected and sent for rework. If 25% of the jobs need rework but never more than once. If 25% of the jobs need rework but reworked jobs are no different in quality than ordinary jobs. 0.75 0.25 A (10) B C I (4)
173
Multiple Paths It is common that there are alternative routes through the process For example: jobs can be split in “fast track”and normal jobs Assume that m different paths originate from a decision point pi = The probability that a job is routed to path i Ti = The time to go down path i CT = p1T1+p2T2+…+pmTm=
174
Example – Processes with Multiple Paths
Consider a process segment consisting of 3 activities A, B & C with activity times 10,15 & 20 minutes respectively On average 20% of the jobs are routed via B and 80% go straight to activity C. What is the average cycle time? 0.8 0.2 A (10) B (15) C (20)
175
Processes with Parallel Activities
If two activities related to the same job are done in parallel the contribution to the cycle time for the job is the maximum of the two activity times. Assuming M process segments in parallel Ti = Average process time for process segment i to be completed CTparallel = Max{T1, T2,…, TM}
176
Example – Cycle Time Analysis of Parallel Activities
Consider a process segment with 5 activities A, B, C, D & E with average activity times: 12, 14, 20, 18 & 15 minutes What is the average cycle time for the process segment? A (12) B (14) C (20) D (18) E (15)
177
Cycle Time Efficiency Measured as the percentage of the total cycle time spent on value adding activities. Theoretical Cycle Time = the cycle time which we would have if only value adding activities were performed That is if the activity times, which include waiting times, are replaced by the processing times See example – Cycle time analysis Exercise 9 & 10, Laguna & Marklund Chapter 5 Cycle Time Efficiency = After this slide it is suitable with a larger example of Cycle Time Analysis, for example, Problems 9 & 10, Chapter 4 in Laguna.
178
Capacity Analysis Focus on assessing the capacity needs and resource utilization in the process Determine the number of jobs flowing through different process segments Determine capacity requirements and utilization based on the flows obtained in 1. The capacity requirements are directly affected by the process configuration Flowcharts are valuable tools Special features to watch out for Rework Multiple Paths Parallel Activities Complements the cycle time analysis!
179
The Effect of Rework on Process Flows
A rework loop implies an increase of the flow rate for that process segment Definitions N = Number of jobs flowing through the rework loop n = Number of jobs arriving to the rework loop from other parts of the process r = Probability that a job needs rework Assuming a job is never reworked more than once Assuming a reworked job is no different than a regular job N = (1+r)n N = n/(1-r)
180
Example – Capacity Analysis with Rework
125 jobs 0.75 0.25 A B C I 100 jobs N = (1+r)n = (1+0.25)100 = 125
181
Multiple Paths and Parallel Activities
Multiple Paths and process flows The flow along a certain path depends on The number of jobs entering the process as a whole (n) The probability for a job to go along a certain path Defining Ni = number of jobs taking path i pi = Probability that a job goes along path i Parallel Activities and process flows All jobs still have to go through all activities if they are in parallel or sequential does not affect the number of jobs flowing through a particular activity Ni = n·pi
182
Analyzing Capacity Needs and Utilization (I)
Need to know Processing times for all activities The type of resource required to perform the activity The number of jobs flowing through each activity The number of available resources of each type Step 1 – Calculate unit load for each resource The total resource time required to process one job Ni = Number of jobs flowing through activity i for every new job entering the process Ti = The processing time for activity i in the current resource M = Total number of activities using the resource Unit load for resource j =
183
Analyzing Capacity Needs and Utilization (II)
Step 2 – Calculate the unit capacity The number of jobs per time unit that can be processed Step 3 – Determine the resource pool capacity A resource pool is a set of identical resources available for use Pool capacity is the number of jobs per time unit that can be processed Let M = Number of resources in the pool Unit capacity for resource j = 1/Unit load for resource j Pool capacity = MUnit capacity = M/unit load
184
Analyzing Capacity Needs and Utilization (III)
Capacity is related to resources not to activities! The process capacity is determined by the bottleneck The bottleneck is the resource or resource pool with the smallest capacity (the slowest resource in terms of jobs/time unit) The slowest resource will limit the process throughput Capacity Utilization The theoretical process capacity is obtained by focusing on processing times as opposed to activity times Delays and waiting times are disregarded The actual process throughput The theoretical capacity! After this slide it is suitable with a larger example of Capacity analysis, for example, Problem 11, Chapter 4 in Laguna. Capacity Utilization =
185
Cycle time Reduction Cycle time and capacity analysis provide valuable information about process performance Helps identify problems Increases process understanding Useful for assessing the effect of design changes Ways of reducing cycle times through process redesign Eliminate activities Reduce waiting and processing time Eliminate rework Perform activities in parallel Move processing time to activities not on the critical path Reduce setup times and enable batch size reduction
186
Example – Critical Activity Reduction
Consider a process with three sequences or paths By moving 2 minutes of activity time from path 2 to path 1 the cycle time is reduced by 2 minutes to CT=45 minutes A B C D E 12 15 18 20 14 Sequence (Path) Time required (minutes) 1. AB E = 41 2. AC E = 47 = CT 3. A D E = 45 Critical path
187
Increasing Process Capacity
Two fundamental ways of increasing process capacity Add resource capacity at the bottleneck Additional equipment, labor or overtime Automation Reduce bottleneck workload Process redesign Shifting activities from the bottleneck to other resources Reducing activity time for bottleneck jobs When the goal is to reduce cycle time and increase capacity careful attention must be given to The resource availability The assignment of activities to resources See also example 5.15 in Laguna & Marklund
188
Theory of Constraints (TOC) (I)
An approach for identifying and managing bottlenecks To increase process flow and thereby process efficiency TOC is focusing on improving the bottom line through Increasing throughput Reducing inventory Reducing operating costs Need operating policies that move the variables in the right directions without violating the given constraints Three broad constraint categories Resource constraints Market constraints Policy constraints
189
Theory of Constraints (TOC) (II)
TOC Methodology Identify the system’s constraints Determine how to exploit the constraints Choose decision/ranking rules for processing jobs in bottleneck Subordinate everything to the decisions in step 2 Elevate the constraints to improve performance For example, increasing bottleneck capacity through investments in new equipment or labor If the current constraints are eliminated return to step 1 Don’t loose inertia, continuous improvement is necessary! See example 5.18 , Chapter 5 in Laguna & Marklund
190
Example – Applying the TOC Methodology
Consider a process with 9 activities and three resource types. Activities 1, 2 & 3 require 10 minutes of processing and the other activities 5 minutes each. There are 3 jobs, following different paths being processed Activities 1, 2 & 3 utilize resource X, activities 4, 5, & 6 resource Y and activities 7, 8 & 9 resource Z. Each resource have 2400 minutes of weekly processing time available Job Routing Demand (Units/week) Profit Margin A 4, 8, and 9 50 20 B 1, 2, 3, 5, 6, 7, and 8 100 75 C 2, 3, 4, 5, 6, 7, 8, and 9 60
191
Introduction to Queuing and Simulation
Chapter 6 Business Process Modeling, Simulation and Design
192
Overview (I) What is queuing/ queuing theory?
Why is it an important tool? Examples of different queuing systems Components of a queuing system The exponential distribution & queuing Stochastic processes Some definitions The Poisson process Terminology and notation Little’s formula Birth and Death Processes
193
Overview (II) Important queuing models with FIFO discipline
The M/M/1 model The M/M/c model The M/M/c/K model (limited queuing capacity) The M/M/c//N model (limited calling population) Priority-discipline queuing models Application of Queuing Theory to system design and decision making
194
Overview (III) Simulation – What is that? Building a simulation model
Why is it an important tool? Building a simulation model Discrete event simulation Structure of a BPD simulation project Model verification and validation Example – Simulation of a M/M/1 Queue
195
What is Queuing Theory? Mathematical analysis of queues and waiting times in stochastic systems. Used extensively to analyze production and service processes exhibiting random variability in market demand (arrival times) and service times. Queues arise when the short term demand for service exceeds the capacity Most often caused by random variation in service times and the times between customer arrivals. If long term demand for service > capacity the queue will explode!
196
Why is Queuing Analysis Important?
Capacity problems are very common in industry and one of the main drivers of process redesign Need to balance the cost of increased capacity against the gains of increased productivity and service Queuing and waiting time analysis is particularly important in service systems Large costs of waiting and of lost sales due to waiting Prototype Example – ER at County Hospital Patients arrive by ambulance or by their own accord One doctor is always on duty More and more patients seeks help longer waiting times Question: Should another MD position be instated?
197
A Cost/Capacity Tradeoff Model
Process capacity Cost Cost of waiting Cost of service Total cost
198
Examples of Real World Queuing Systems?
Commercial Queuing Systems Commercial organizations serving external customers Ex. Dentist, bank, ATM, gas stations, plumber, garage … Transportation service systems Vehicles are customers or servers Ex. Vehicles waiting at toll stations and traffic lights, trucks or ships waiting to be loaded, taxi cabs, fire engines, elevators, buses … Business-internal service systems Customers receiving service are internal to the organization providing the service Ex. Inspection stations, conveyor belts, computer support … Social service systems Ex. Judicial process, the ER at a hospital, waiting lists for organ transplants or student dorm rooms …
199
Components of a Basic Queuing Process
Input Source The Queuing System Served Jobs Calling Population Jobs Service Mechanism Queue leave the system Arrival Process Queue Discipline Service Process Queue Configuration
200
Components of a Basic Queuing Process (II)
The calling population The population from which customers/jobs originate The size can be finite or infinite (the latter is most common) Can be homogeneous (only one type of customers/ jobs) or heterogeneous (several different kinds of customers/jobs) The Arrival Process Determines how, when and where customer/jobs arrive to the system Important characteristic is the customers’/jobs’ inter-arrival times To correctly specify the arrival process requires data collection of interarrival times and statistical analysis.
201
Components of a Basic Queuing Process (III)
The queue configuration Specifies the number of queues Single or multiple lines to a number of service stations Their location Their effect on customer behavior Balking and reneging Their maximum size (# of jobs the queue can hold) Distinction between infinite and finite capacity
202
Example – Two Queue Configurations
Servers Multiple Queues Servers Single Queue
203
Multiple v.s. Single Customer Queue Configuration
Multiple Line Advantages Single Line Advantages The service provided can be differentiated Ex. Supermarket express lanes Labor specialization possible Customer has more flexibility Balking behavior may be deterred Several medium-length lines are less intimidating than one very long line Guarantees fairness FIFO applied to all arrivals No customer anxiety regarding choice of queue Avoids “cutting in” problems The most efficient set up for minimizing time in the queue Jockeying (line switching) is avoided
204
Components of a Basic Queuing Process (IV)
The Service Mechanism Can involve one or several service facilities with one or several parallel service channels (servers) - Specification is required The service provided by a server is characterized by its service time Specification is required and typically involves data gathering and statistical analysis. Most analytical queuing models are based on the assumption of exponentially distributed service times, with some generalizations. The queue discipline Specifies the order by which jobs in the queue are being served. Most commonly used principle is FIFO. Other rules are, for example, LIFO, SPT, EDD… Can entail prioritization based on customer type.
205
Mitigating Effects of Long Queues
Concealing the queue from arriving customers Ex. Restaurants divert people to the bar or use pagers, amusement parks require people to buy tickets outside the park, banks broadcast news on TV at various stations along the queue, casinos snake night club queues through slot machine areas. Use the customer as a resource Ex. Patient filling out medical history form while waiting for physician Making the customer’s wait comfortable and distracting their attention Ex. Complementary drinks at restaurants, computer games, internet stations, food courts, shops, etc. at airports Explain reason for the wait Provide pessimistic estimates of the remaining wait time Wait seems shorter if a time estimate is given. Be fair and open about the queuing disciplines used
206
A Commonly Seen Queuing Model (I)
The Queuing System The Service Facility C S = Server C S • The Queue Customers (C) C C C … C Customer =C
207
A Commonly Seen Queuing Model (II)
Service times as well as interarrival times are assumed independent and identically distributed If not otherwise specified Commonly used notation principle: A/B/C A = The interarrival time distribution B = The service time distribution C = The number of parallel servers Commonly used distributions M = Markovian (exponential) - Memoryless D = Deterministic distribution G = General distribution Example: M/M/c Queuing system with exponentially distributed service and inter-arrival times and c servers
208
The Exponential Distribution and Queuing
The most commonly used queuing models are based on the assumption of exponentially distributed service times and interarrival times. Definition: A stochastic (or random) variable Texp( ), i.e., is exponentially distributed with parameter , if its frequency function is: The Cumulative Distribution Function is: The mean = E[T] = 1/ The Variance = Var[T] = 1/ 2
209
The Exponential Distribution
fT(t) Probability density t Mean= E[T]=1/ Time between arrivals
210
Properties of the Exp-distribution (I)
Property 1: fT(t) is strictly decreasing in t P(0Tt) > P(t T t+t) for all t, t0 Implications Many realizations of T (i.e.,values of t) will be small; between zero and the mean Not suitable for describing the service time of standardized operations when all times should be centered around the mean Ex. Machine processing time in manufacturing Often reasonable in service situations when different customers require different types of service Often a reasonable description of the time between customer arrivals
211
Properties of the Exp-distribution (II)
Property 2: Lack of memory P(T>t+t | T>t) = P(T >t) for all t, t0 Implications It does not matter when the last customer arrived, (or how long service time the last job required) the distribution of the time until the next one arrives (or the distribution of the next service time) is always the same. Usually a fair assumption for interarrival times For service times, this can be more questionable. However, it is definitely reasonable if different customers/jobs require different service
212
Properties of the Exp-distribution (III)
Property 3: The minimum of independent exponentially distributed random variables is exponentially distributed Assume that {T1, T2, …, Tn} represent n independent and exponentially distributed stochastic variables with parameters {1, 2, …, n}. Let U=min {T1, T2, …, Tn} Implications Arrivals with exponentially distributed interarrival times from n different input sources with arrival intensities {1, 2, …, n} can be treated as a homogeneous process with exponentially distributed interarrival times of intensity = 1+ 2+…+ n. Service facilities with n occupied servers in parallel and service intensities {1, 2, …, n}can be treated as one server with service intensity = 1+2+…+n
213
Properties of the Exp-distribution (IV)
Relationship to the Poisson distribution and the Poisson Process Let X(t) be the number of events occurring in the interval [0,t]. If the time between consecutive events is T and Texp() X(t)Po(t) {X(t), t0} constitutes a Poisson Process
214
Stochastic Processes in Continuous Time
Definition: A stochastic process in continuous time is a family {X(t)} of stochastic variables defined over a continuous set of t-values. Example: The number of phone calls connected through a switch board Definition: A stochastic process {X(t)} is said to have independent increments if for all disjoint intervals (ti, ti+hi) the differences Xi(ti+hi)Xi(ti) are mutually independent. X(t)=# Calls t
215
The Poisson Process The standard assumption in many queuing models is that the arrival process is Poisson Two equivalent definitions of the Poisson Process The times between arrivals are independent, identically distributed and exponential X(t) is a Poisson process with arrival rate iff. X(t) have independent increments b) For a small time interval h it holds that P(exactly 1 event occurs in the interval [t, t+h]) = h + o(h) P(more than 1 event occurs in the interval [t, t+h]) = o(h)
216
Properties of the Poisson Process
Poisson processes can be aggregated or disaggregated and the resulting processes are also Poisson processes a) Aggregation of N Poisson processes with intensities {1, 2, …, n} renders a new Poisson process with intensity = 1+ 2+…+ n. b) Disaggregating a Poisson process X(t)Po(t) into N sub-processes {X1(t), X2(t), , …, X3(t)} (for example N customer types) where Xi(t) Po(it) can be done if – For every arrival the probability of belonging to sub-process i = pi – p1+ p2+…+ pN = 1, and i = pi
217
Illustration – Disaggregating a
Poisson Process p1 X(t)Po(t) p2 pN
218
Terminology and Notation
The state of the system = the number of customers in the system Queue length = (The state of the system) – (number of customers being served) N(t) = Number of customers/jobs in the system at time t Pn(t) = The probability that at time t, there are n customers/jobs in the system. n = Average arrival intensity (= # arrivals per time unit) at n customers/jobs in the system n = Average service intensity for the system when there are n customers/jobs in it. (Note, the total service intensity for all occupied servers) = The utilization factor for the service facility. (= The expected fraction of the time that the service facility is being used)
219
Example – Service Utilization Factor
Consider an M/M/1 queue with arrival rate = and service intensity = = Expected capacity demand per time unit = Expected capacity per time unit Similarly if there are c servers in parallel, i.e., an M/M/c system but the expected capacity per time unit is then c*
220
Queuing Theory Focus on Steady State
Steady State condition Enough time has passed for the system state to be independent of the initial state as well as the elapsed time The probability distribution of the state of the system remains the same over time (is stationary). Transient condition Prevalent when a queuing system has recently begun operations The state of the system is greatly affected by the initial state and by the time elapsed since operations started The probability distribution of the state of the system changes with time With few exceptions Queuing Theory has focused on analyzing steady state behavior
221
Transient and Steady State Conditions
Illustration of transient and steady-state conditions N(t) = number of customers in the system at time t, E[N(t)] = represents the expected number of customers in the system.
222
Notation For Steady State Analysis
Pn = The probability that there are exactly n customers/jobs in the system (in steady state, i.e., when t) L = Expected number of customers in the system (in steady state) Lq = Expected number of customers in the queue (in steady state) W = Expected time a job spends in the system Wq= Expected time a job spends in the queue
223
Little’s Formula Revisited
Assume that n = and n = for all n Assume that n is dependent on n
224
Birth-and-Death Processes
The foundation of many of the most commonly used queuing models Birth – equivalent to the arrival of a customer or job Death – equivalent to the departure of a served customer or job Assumptions Given N(t)=n, The time until the next birth (TB) is exponentially distributed with parameter n (Customers arrive according to a Po-process) The remaining service time (TD) is exponentially distributed with parameter n TB & TD are mutually independent stochastic variables and state transitions occur through exactly one Birth (n n+1) or one Death (n n–1)
225
A Birth-and-Death Process Rate Diagram
Excellent tool for describing the mechanics of a Birth-and-Death process 0 1 n-1 n 1 2 n-1 n n+1 1 2 n n+1 n = State n, i.e., the case of n customers/jobs in the system
226
Steady State Analysis of B-D Processes (I)
In steady state the following balance equation must hold for every state n (proved via differential equations) The Rate In = Rate Out Principle: Mean entrance rate = Mean departure rate In addition the probability of being in one of the states must equal 1
227
Steady State Analysis of B-D Processes (II)
Balance Equation 1 n C0 C2
228
Steady State Analysis of B-D Processes (III)
Steady State Probabilities Expected Number of Jobs in the System and in the Queue Assuming c parallel servers
229
The M/M/1 - model Assumptions - the Basic Queuing Process
Infinite Calling Populations Independence between arrivals The arrival process is Poisson with an expected arrival rate Independent of the number of customers currently in the system The queue configuration is a single queue with possibly infinite length No reneging or balking The queue discipline is FIFO The service mechanism consists of a single server with exponentially distributed service times = expected service rate when the server is busy
230
The M/M/1 Model n= and n = for all values of n=0, 1, 2, …
1 n n-1 2 n+1 Steady State condition: = (/) < 1 Pn = n(1- ) P0 = 1- P(nk) = k L=/(1- ) Lq= 2/(1- ) = L- W=L/=1/(- ) Wq=Lq/= /( (- ))
231
The M/M/c Model (I) Generalization of the M/M/1 model Steady State
2 (c-1) c 1 c c-2 2 c+1 c-1 (c-2) Generalization of the M/M/1 model Allows for c identical servers working independently from each other Steady State Condition: =(/c)<1
232
The M/M/c Model (II) A Condition for existence of a steady state solution is that = /(c) <1 Little’s Formula Wq=Lq/ W=Wq+(1/) Little’s Formula L=W= (Wq+1/ ) = Lq+ /
233
Example – ER at County Hospital
Situation Patients arrive according to a Poisson process with intensity ( the time between arrivals is exp() distributed. The service time (the doctor’s examination and treatment time of a patient) follows an exponential distribution with mean 1/ (=exp() distributed) The ER can be modeled as an M/M/c system where c=the number of doctors Data gathering = 2 patients per hour = 3 patients per hour Questions Should the capacity be increased from 1 to 2 doctors? How are the characteristics of the system (, Wq, W, Lq and L) affected by an increase in service capacity?
234
Summary of Results – County Hospital
Interpretation To be in the queue = to be in the waiting room To be in the system = to be in the ER (waiting or under treatment) Is it warranted to hire a second doctor ? Characteristic One doctor (c=1) Two Doctors (c=2) 2/3 1/3 P0 1/2 (1-P0) P1 2/9 Lq 4/3 patients 1/12 patients L 2 patients 3/4 patients Wq 2/3 h = 40 minutes 1/24 h = 2.5 minutes W 1 h 3/8 h = 22.5 minutes
235
The M/M/c/K – Model (I) An M/M/c model with a maximum of K customers/jobs allowed in the system If the system is full when a job arrives it is denied entrance to the system and the queue. Interpretations A waiting room with limited capacity (for example, the ER at County Hospital), a telephone queue or switchboard of restricted size Customers that arrive when there is more than K clients/jobs in the system choose another alternative because the queue is too long (Balking)
236
The M/M/c/K – Model (II)
Still a Birth-and-Death process but with a state dependent arrival intensity Observation The M/M/c/K model always has a steady state solution since the queue can never “explode”
237
The M/M/c/K – Model (III)
The state diagram has exactly K states provided that c<K The general expressions for the steady state probabilities, waiting times, queue lengths etc. are obtained through the balance equations as before (Rate In = Rate Out; for every state) 2 (c-1) c 1 K-1 c-1 2 K c 3
238
Results for the M/M/1/K – Model
Where
239
The M/M/c//N – Model (I)
An M/M/c model with limited calling population, i.e., N clients A common application: Machine maintenance c service technicians is responsible for keeping N service stations (machines) running, that is, to repair them as soon as they break Customer/job arrivals = machine breakdowns Note, the maximum number of clients in the system = N Assume that (N-n) machines are operating and the time until breakdown for each machine i, Ti, is exponentially distributed (Tiexp()). If U = the time until the next breakdown U = Min{T1, T2, …, TN-n} Uexp((N-n))).
240
The M/M/c//N – Model (II)
The State Diagram (c service technicians and N machines) = Arrival intensity per operating machine = The service intensity for a service technician General expressions for this queuing model can be obtained from the balance equations as before N (N-1) (N-(c-1)) 2 (c-1) c 1 N-1 c-1 2 N c 3
241
Priority-Discipline Queuing Models
For situations where different customers have different priorities For example, ER operations, VIP customers at nightclubs… Assuming a situation with N priority classes (where class 1 has the highest priority) there are two fundamental priority principles to consider. Non-Preemptive priorities A customer being served cannot be ejected back into the queue to leave place for a customer with higher priority Preemptive priorities A customer of lower priority that is being served will be thrown back into the queue to leave room for a higher priority customer Assuming that all customers experience independent exp() service times and arrive according to Poisson processes both models can be analyzed as special case M/M/c models
242
Queuing Modeling and System Design (I)
Design of queuing systems usually involve some kind of capacity decision The number of service stations The number of servers per station The service time for individual servers The corresponding decision variables are , c and Examples: The number of doctors in a hospital, The number of exits and cashiers in a supermarket, The choice of machine type at a new investment decision, The localization of toilets in a new building, etc…
243
Queuing Modeling and System Design (II)
Two fundamental questions when designing (queuing) systems Which service level should we aim for? How much capacity should we acquire? The cost of increased capacity must be balanced against the cost reduction due to shorter waiting time Specify a waiting cost or a shortage cost accruing when customers have to wait for service or… … Specify an acceptable service level and minimize the capacity under this condition The shortage or waiting cost rate is situation dependent and often difficult to quantify Should reflect the monetary impact a delay has on the organization where the queuing system resides
244
Different Shortage Cost Situations
External customers arrive to the system Profit organizations The shortage cost is primarily related to lost revenues – “Bad Will” • Non-profit organizations The shortage cost is related to a societal cost Internal customers arrive to the system The shortage cost is related to productivity loss and associated profit loss Usually it is easier to estimate the shortage costs in situation 2. than in situation 1.
245
Analyzing Design-Cost Tradeoffs
Given a specified shortage or waiting cost function the analysis is straightforward Define WC = Expected Waiting Cost (shortage cost) per time unit SC = Expected Service Cost (capacity cost) per time unit TC = Expected Total system cost per time unit The objective is to minimize the total expected system cost TC Cost Min TC = WC + SC SC WC Process capacity
246
Analyzing Linear Waiting Costs
Expected Waiting Costs as a function of the number of customers in the system Cw = Waiting cost per customer and time unit CwN = Waiting cost per time unit when N customers in the system Expected Waiting Costs as a function of the number of customers in the queue
247
Analyzing Service Costs
The expected service costs per time unit, SC, depend on the number of servers and their speed Definitions c = Number of servers = Average server intensity (average time to serve one customer) CS() = Expected cost per server and time unit as a function of SC = c*CS()
248
A Decision Model for System Design
Determining and c Both the number of servers and their speed can be varied Usually only a few alternatives are available Definitions A = The set of available - options Optimization Enumerate all interesting combinations of and c, compute TC and choose the cheapest alternative From a structural point of view, a few fast servers are usually better than several slow ones with the same maximum capacity
249
Example – “Computer Procurement”
A university is about to lease a super computer There are two alternatives available The M computer which is more expensive to lease but also faster The C computer which is cheaper but slower Processing times and times between job arrivals are exponential M/M/1 model = 20 jobs per day M = 30 jobs per day C = 25 jobs per day The leasing and waiting costs: Leasing price: CM = $500 per day, CC = $350 per day The waiting cost per job and time unit job is estimated to $50 per job and day Question: Which computer should the university choose in order to minimize the expected costs?
250
Simulation – What is it? Experiment with a model mimicking the real world system Ex. Flight simulation, wind tunnels, … In BPD situations computer based simulation is used for analyzing and evaluating complex stochastic systems Uncertain service and inter-arrival times
251
Simulation – Why use it? Cheaper and less risky than experimenting with the actual system. Stimulates creativity since it is easy to test the effect of new ideas A powerful complement to the traditional symbolical and analytical tools Fun tool to work with!
252
Simulation v.s. Symbolic & Analytical Tools
Strengths Provides a quantitative measure Flexible – can handle any kind of complex system or statistical interdependencies Capable of finding inefficiencies otherwise not detected until the system is in operation Weaknesses Can take a long time to build Usually requires a substantial amount of data gathering Easy to misrepresent reality and draw faulty conclusions Generally not suitable for optimizing system parameters A simulation model is primarily descriptive while an optimization model is by nature prescriptive
253
Modern Simulation Software Packages are Breaking Compromises
Graphical interfaces Achieves the descriptive benefits of symbolic tools like flow charts Optimization Engines Enables efficient automated search for best parameter values
254
Building a Simulation Model
General Principles The system is broken down into suitable components or entities The entities are modeled separately and are then connected to a model describing the overall system A bottom-up approach! The basic principles apply to all types of simulation models Static or Dynamic Deterministic or Stochastic Discrete or continuous In BPD and OM situations computer based Stochastic Discrete Event Simulation (e.g. in Extend) is the natural choice Focuses on events affecting the state of the system and skips all intervals in between
255
Steps in a BPD Simulation Project
1. Problem formulation 2. Set objectives and overall project plan Phase 1 Problem Definition 3. Model conceptualization 4. Data Collection 5. Model Translation 6. Verified 7. Validated Yes No Phase 2 Model Building Phase 3 Experimentation 11. Documentation, reporting and implementation Phase 4 Implementation 8. Experimental Design 9. Model runs and analysis 10. More runs No Yes
256
Model Verification and Validation
Verification (efficiency) Is the model correctly built/programmed? Is it doing what it is intended to do? Validation (effectiveness) Is the right model built? Does the model adequately describe the reality you want to model? Does the involved decision makers trust the model? Two of the most important and most challenging issues in performing a simulation study
257
Model Verification Methods
Find alternative ways of describing/evaluating the system and compare the results Simplification enables testing of special cases with predictable outcomes Removing variability to make the model deterministic Removing multiple job types, running the model with one job type at a time Reducing labor pool sizes to one worker Build the model in stages/modules and incrementally test each module Uncouple interacting sub-processes and run them separately Test the model after each new feature that is added Simple animation is often a good first step to see if things are working as intended
258
Validation - an Iterative Calibration Process
The Real System Calibration and Validation Conceptual validation Conceptual Model Assumptions on system components Structural assumptions which define the interactions between system components 3. Input parameters and data assumptions Operational Model (Computerized representation) Model verification
259
Example – Simulation of a M/M/1 Queue
Assume a small branch office of a local bank with only one teller. Empirical data gathering indicates that inter-arrival and service times are exponentially distributed. The average arrival rate = = 5 customers per hour The average service rate = = 6 customers per hour Using our knowledge of queuing theory we obtain = the server utilization = 5/6 0.83 Lq = the average number of people waiting in line Wq = the average time spent waiting in line Lq = 0.832/(1-0.83) 4.2 Wq = Lq/ 4.2/5 0.83 How do we go about simulating this system? How do the simulation results match the analytical ones?
260
Introduction to Extend
Chapter 7 Business Process Modeling, Simulation and Design
261
Introduction to Extend
NOTE: We recommend presenting this chapter by running Extend 6.0 directly, and interactively show how the program works. However, for your convenience, we have attached a selection of the figures/screenshots from Chapter 7 of the book as the basis for an in class presentation without access to a computer with Extend installed.
262
Extend Elements Executive Import Repository Operation Export
Basic Extend Blocks
263
Extend Elements A Simple credit application process
264
Tutorial: Basic Queuing Model
Main Extend window
265
Basic Queuing Model Import dialog window
266
Basic Queuing Model Underwriting process model with a single team
267
Data Collection and Statistics
Blocks in the Statistics Submenu of the Discrete Event library
268
Queue Statistics Stack block
269
Queue Statistics Underwriting process model with a stack block
270
Queue Statistics Queue statistics for “In Box” block
271
Queue Statistics Plotter, Discrete Event block
272
Queue Statistics Underwriting process model with a plotter block
273
Queue Statistics Waiting time vs. simulation time
274
Random Processing Times
Input Random Number block
275
Random Processing Times
Underwriting process model with random processing times
276
Random Processing Times
Queue statistics for “In Box” block (100-week run)
277
Random Processing Times
Dialog window of the Underwriting block
278
Random Processing Times
Average utilization of the underwriting team
279
Adding a Second Team Underwriting process model with two teams
280
Adding a Second Team Transaction block
281
Adding a Second Team Modeling multiple teams with a transition block
282
Adding a Labor Pool Additional blocks from the BPR and Discrete Event libraries
283
Adding a Labor Pool Underwriting process model with a review activity and a labor pool
284
Adding a Labor Pool “Approve?” block dialog
285
Animating the Model Animate tab of “Requests In” block
286
Customizing the Animation
Animate tab of “Teams” block
287
Calculating Costs Cost tab of labor pool block
288
Calculating Costs Cost per request
289
Calculating Costs Unbatch block dialog
290
Calculating Costs Underwriting process model with cost collection blocks
291
Using the Flowchart View
Flowchart view of the simulation model for the Underwriting process
292
Exercise 3 Process with two servers in series
293
Exercise 4 1 2 3 4 5 Teller configuration (multiple queues)
294
Modeling and Simulating Business Processes
Chapter 8 Business Process Modeling, Simulation and Design
295
Simulating Business Processes
NOTE: We recommend presenting this chapter by running Extend 6.0 directly, and interactively show how the program works. However, for your convenience, we have attached a selection of the figures/screenshots from Chapter 8 of the book as the basis for an in class presentation without access to a computer with Extend installed.
296
Generating Items Import block dialog showing and exponential inter-arrival times with mean of 6 minutes
297
Generating Items Input Data block used to change the first parameter of the inter-arrival time distribution in the Import block
298
Generating Items Input Data block dialog for dry cleaner example
299
Generating Items Program block (and dialog window) connected to a Stack block
300
Prioritizing Items Selection of a priority queue in the Queue tab of a Stack block
301
Balking Model of a single server with a queue, where customers
balk if the line reaches a specified number of customers
302
Balking Dialog window of the Decision(2) block
303
Reneging Model of a single server with a queue, where customers hang up after being on hold for specified amount of time
304
Reneging Stack block to simulate a reneging queue with reneging time of 5 minutes
305
Priority Queues Admissions process with a priority queue that allows
patients to go in front of the line after filling out additional forms
306
Priority Queues Dialog window of the Stack block
307
Routing - Multiple Paths
0.20 0.30 0.50 Incoming job Example of probabilistic routing of an incoming job
308
Multiple Paths Illustration of probabilistic routing with Extend
309
Multiple Paths Illustration of tactical routing with customers choosing the shortest line
310
Routing – Parallel Paths
Prepare invoice Assemble order Receive order Ship order Parallel activities in an order fulfillment process
311
Parallel Paths Operation, Reverse block and dialog window
312
Parallel Paths Operation block that batches two items
313
Parallel Paths Extend model of the order fulfillment process
314
Processing Time Slider control
315
Processing Time Input Data block to model variable processing time
316
Processing Time Processing time based on the value of the “ProcTime” attribute
317
Batching Batch block and dialog window
318
Batching Preserving uniqueness when batching a purchase order and an agent
319
Batching Unbatching items with an Unbatch block
320
Batching – Resources Model that batches patients and nurses
321
Activity-based Costing
Cost tab of the Import block
322
Activity-based Costing
Model to accumulate cost per unpacked crate
323
Activity-based Costing
Dialog of the Cost by Item block
324
Cycle Time Analysis Two operations in series with a Timer block to measure cycle time
325
Cycle Time Analysis Histogram of cycle times and average cycle time vs. simulation time
326
Model Enhancements Slider control to set the mean value of an Exponential distribution
327
Model Enhancements Meter connected to the utilization output of Labor Pool block
328
Model Enhancements Clone layer tool
329
Case: Software Support
Engineer Incoming calls Real-time response s Problems resolved Documented software support process
330
Case: Software Support
Simulation model of actual process
331
Case: Software Support
Simulation model of documented process
332
Case: Hospital Admissions
Schematic representation of the hospital admissions process
333
Case: Hospital Admissions
Extend model of the current admissions process
334
Case: Hospital Admissions
Arrivals block from the admissions model
335
Case: Hospital Admissions
Admissions block from the admissions model
336
Case: Hospital Admissions
Registration and Lab block from the admissions model
337
Case: Hospital Admissions
Rooms block from the admissions model
338
Case: Hospital Admissions
Histogram of cycle times for type 1 patients
339
Case: Hospital Admissions
Extend model of the redesigned admissions process
340
Exercise 1 Measuring cycle times of different types of jobs A B C D E
Type I Types II & III Measuring cycle times of different types of jobs
341
Exercise 2 Investigating the effect of pooling resources Type 1 Type 2
B1 A2 B2 A3 B3 Type 1 C1 C2 C3 Type 2 Type 3 Investigating the effect of pooling resources
342
Exercise 4 Assessing process performance
343
Exercise 7 A B C E D F 0.8 0.2 Priority queues
344
Exercise 8 Flowchart for Exercise 8
345
Project: X-ray Process
1 2 3 4 5 6 7 10 11 12 8 9 25% Flowchart for the X-ray process
346
Project: An Assembly Factory
2 1 3 4 5 Storage of inbound material Parallel operations (Workstations 1-3) Assembly (Workstation 4) Painting (Workstation 5) Inspection Rework Flowchart of the production process for the Assembly Factory
347
Simulation Input and Output Data Analysis
Chapter 9 Business Process Modeling, Simulation and Design Augmented with material from other sources
348
Overview Analysis of input data Analysis of Output Data
Identification of field data distributions Goodness-of-fit tests Random number generation Analysis of Output Data Non-terminating v.s. terminating processes Confidence intervals Hypothesis testing for comparing designs
349
Why Input and Output Data Analysis?
Simulation Model Output Data Input Data Random Analysis of input data Necessary for building a valid model Three aspects Identification of (time) distributions Random number generation Generation of random variates Integrated into Extend Analysis of output data Necessary for drawing correct conclusions The reported performance measures are typically random variables!
350
Capturing Randomness in Input Data
Collect raw field data and use as input for the simulation No question about relevance Expensive/impossible to retrieve a large enough data set Not available for new processes Not available for multiple scenarios No sensitivity analysis Very valuable for model validation Generate artificial data to use as input data Must capture the characteristics of the real data Collect a sufficient sample of field data Characterize the data statistically – Distribution type and parameters Generate random artificial data mimicking the real data High flexibility – easy to handle new scenarios Cheap Requires proper statistical analysis to ensure model validity
351
Procedure for Modeling Input Data
1. Gather data from the real system Plot histograms of the data Compare the histogram graphically (“eye-balling”) with shapes of well known distribution functions How about the tails of the distribution, limited or unlimited? How to handle negative outcomes? 2. Identify an appropriate distribution family Distribution hypothesis rejected 3. Estimate distribution parameters and pick an “exact” distribution Informal test – “eye-balling” Formal tests, for example 2 - test Kolmogorov-Smirnov test 4. Perform Goodness–of–fit test (Reject the hypothesis that the picked distribution is correct?) If a known distribution can not be accepted Use an empirical distribution
352
Example – Modeling Interarrival Times (I)
Data gathering from the real system Interarrival Time (t) Frequency 0t<3 23 3t<6 10 6t<9 5 9t<12 1 12t<15 15t<18 2 18t<21 21t<24 24t<27 Etc.
353
Example – Modeling Interarrival Times (II)
Identify an appropriate distribution type/family Plot a histogram Divide the data material into appropriate intervals Usually of equal size Determine the event frequency for each interval (or bin) Plot the frequency (y-axis) for each interval (x-axis) The Exponential distribution seems to be a good first guess!
354
Example – Modeling Interarrival Times (III)
Estimate the parameters defining the chosen distribution In the current example Exp()has been chosen need to estimate the parameter ti = the ith interarrival time in the collected sample of n observations
355
Example – Modeling Interarrival Times (III)
Perform Goodness-of-fit test The purpose is to test the hypothesis that the data material is adequately described by the “exact” distribution chosen in steps 1-3. Two of the most well known standardized tests are The 2-test Should not be applied if the sample size n<20 The Kolmogorov-Smirnov test A relatively simple but imprecise test Often used for small sample sizes The 2-test will be applied for the current example
356
Performing a 2-Test (I)
In principle A statistical test comparing the relative frequencies for the intervals/bins in a histogram with the theoretical probabilities of the chosen distribution Assumptions The distribution involves k parameters estimated from the sample The sample contains n observations (sample size=n) F0(x) denotes the chosen/hypothesized CDF Data: x1, x2, …, xn (n observations from the real system) Model: X1, X2,…, Xn (Random variables, independent and identically distributed with CDF F(x)) Null hypothesis H0: F(x) = F0(x) Alternative hypothesis HA: F(x) F0(x)
357
Performing a 2-Test (II)
Take the entire data range and divide it into r non overlapping intervals or bins f0(x) The area = p2 = F0(a2) - F0(a1) Data values Min=a0 a1 a2 ar-1 ar=Max … a3 ar-2 Bin: 1 2 3 r-1 r pi = The probability that an observation X belongs to bin i The Null Hypothesis pi = F0(ai) - F0(ai-1) To improve the accuracy of the test choose the bins (intervals) so that the probabilities pi (i=1,2, …r) are equal for all bins
358
Performing a 2-Test (III)
2. Define r random variables Oi, i=1, 2, …r Oi=number of observations in bin i (= the interval (ai-1, ai]) If H0 is true the expected value of Oi = n*pi Oi is Binomially distributed with parameters n and pi 3. Define the test variable T If H0 is true T follows a 2(r-k-1) distribution T = The critical value of T corresponding to a significance level obtained from a 2(r-k-1) distribution table Tobs = The value of T computed from the data material If Tobs > T H0 can be rejected on the significance level
359
Validity of the 2-Test Depends on the sample size n and on the bin selection (the size of the intervals) Rules of thumb The 2-test is acceptable for ordinary significance levels (=1%, 5%) if the expected number of observations in each interval is greater than 5 (n*pi>5 for all i) In the case of continuous data and a bin selection such that pi is equal for all bins n20 Do not use the 2-test 20<n 50 5-10 bins recommendable 50<n 100 bins recommendable n >100 n0.5 – 0.2n bins recommendable
360
Example – Modeling Interarrival Times (IV)
Hypothesis – the interarrival time Y is Exp(0.084) distributed H0: YExp(0.084) HA: YExp(0.084) Bin sizes are chosen so that the probability pi is equal for all r bins and n*pi>5 for all i Equal pi pi=1/r n*pi>5 n/r > 5 r<n/5 n=50 r<50/5=10 Choose for example r=8 pi=1/8 Determining the interval limits ai, i=0,1,…8 i=1 a1=ln(1-(1/8))/(-0.084)=1.590 i=2 a2=ln(1-(2/8))/(-0.084)=3.425 i=8 a8 =ln(1-(8/8))/(-0.084)=
361
Example – Modeling Interarrival Times (V)
Computing the test statistic Tobs Note: oi = the actual number of observations in bin i Determining the critical value T If H0 is true T2(8-1-1)=2(6) If =0.05 P(T T0.05)=1-=0.95 /2 table/ T0.05=12.60 Rejecting the hypothesis Tobs=39.6>12.6= T0.05 H0 is rejected on the 5% level
362
The Kolmogorov-Smirnov test (I)
Advantages over the chi-square test Does not require decisions about bin ranges Often applied for smaller sample sizes Disadvantages Ideally all distribution parameters should be known with certainty for the test to be valid A modified version based on estimated parameter values exist for the Normal, Exponential and Weibull distributions In practice often used for other distributions as well For samples with n30 the 2-test is more reliable!
363
The Kolmogorov-Smirnov test (II)
Compares an empirical “relative-frequency” CDF with the theoretical CDF (F(x)) of a chosen (hypothesized) distribution The empirical CDF = Fn(x) = (number of xix)/n n=number of observations in the sample xi=the value of the ith smallest observation in the sample Fn(xi)=i/n Procedure Order the sample data from the smallest to the largest value Compute D+ , D– and D = max{D+ , D–} Find the tabulated critical KS value corresponding to the sample size n and the chosen significance level, If the critical KS value D reject the hypothesis that F(x) describes the data material’s distribution
364
Distribution Choice in Absence of Sample Data
Common situation especially when designing new processes Try to draw on expert knowledge from people involved in similar tasks When estimates of interval lengths are available Ex. The service time ranges between 5 and 20 minutes Plausible to use a Uniform distribution with min=5 and max=20 When estimates of the interval and most likely value exist Ex. min=5, max=20, most likely=12 Plausible to use a Triangular distribution with those parameter values When estimates of min=a, most likely=c, max=b and the average value=x-bar are available Use a -distribution with parameters and
365
Random Number Generators
Needed to create artificial input data to the simulation model Generating truly random numbers is difficult Computers use pseudo-random number generators based on mathematical algorithms – not truly random but good enough A popular algorithm is the “linear congruential method” 1. Define a random seed x0 from which the sequence is started 2. The next “random” number in the sequence is obtained from the previous through the relation where a, c, and m are integers > 0
366
Example – The Linear Congruential Method
Assume that m=8, a=5, c=7 and x0=4 n xn 5xn+7 (5xn+7)/8 xn+1 4 27 /8 3 1 22 /8 6 2 37 /8 5 32 /8 7 /8 42 /8 17 /8 12 /8 Larger m longer sequence before it starts repeating itself
367
The Runs Test (I) Test for detecting dependencies in a sequence of generated random numbers A run is defined as a sequence of increasing or decreasing numbers “+” indictes an increasing run “–” indicates a decreasing run Ex. Numbers: 1, 7, 8, 6, 5, 3, 4, 10, 12, 15 runs: – – – The test is based on comparing the number of runs in a true random sequence with the number of runs in the observed sequence
368
The Runs Test (II) Hypothesis: H0: Sequence of numbers is independent
HA: Sequence of numbers is not independent R = # runs in a truly random sequence of n numbers (random variable) Have been shown that… R=(2n-1)/3 R=(16n-29)/90 RN(R, R) Test statistic: Z={(R-R)/R}N(0,1) Assuming: confidence level and a two sided test P(-Z/2ZZ/2)=1- H0 is rejected if Zobserved> Z/2
369
Generating Random Variates
Assume random numbers, r, from a Uniform (0, 1) distribution are available Random numbers from any distribution can be obtained by applying the “inverse transformation technique” The inverse Transformation Technique Generate a U[0, 1] distributed random number r T is a random variable with a CDF FT(t) from which we would like to obtain a sequence of random numbers Note: 0 FT(t) 1 for all values of t t is a random number from the distribution of T, i.e., a realization of T
370
Analysis of Simulation Output Data
The output data collected from a simulation model are realizations of stochastic variables Results from random input data and random processing times Statistical analysis is required to Estimate performance characteristics Mean, variance, confidence intervals etc. for output variables Compare performance characteristics for different designs The validity of the statistical analysis and the design conclusions are contingent on a careful sampling approach Sample sizes – run length and number of runs. Inclusion or exclusion of “warm-up” periods? One long simulation run or several shorter ones?
371
Terminating v.s. Non-Terminating Processes
372
Non-Terminating Processes
Does not end naturally within a particular time horizon Ex. Inventory systems Usually reach steady state after an initial transient period Assumes that the input data is stationary To study the steady state behavior it is vital to determine the duration of the transient period Examine line plots of the output variables To reduce the duration of the transient (=“warm-up) period Initialize the process with appropriate average values
373
Illustration Transient and Steady state
Line plot of cycle times and average cycle time Transient state Steady state
374
Terminating Processes
Ends after a predetermined time span Typically the system starts from an empty state and ends in an empty state Ex. A grocery store, a construction project, … Terminating processes may or may not reach steady state Usually the transient period is of great interest for these processes Output data usually obtained from multiple independent simulation runs The length of a run is determined by the natural termination of the process Each run need a different stream of random numbers The initial state of each run is typically the same
375
Confidence Intervals and Point Estimates
Statistical estimation of measures from a data material are typically done in two ways Point estimates (single values) Confidence intervals (intervals) The confidence level Indicates the probability of not finding the true value within the interval (Type I error) Chosen by the analyst/manager Determinants of confidence interval width The chosen confidence level Lower wider confidence interval The sample size and the standard deviation () Larger sample smaller standard deviation narrower interval
376
Important Point Estimates
In simulation the most commonly used statistics are the mean and standard deviation () From a sample of n observations Point estimate of the mean: Point estimate of :
377
Confidence Interval for Population Means (I)
Characteristics of the point estimate for the population mean Xi = Random variable representing the value of the ith observation in a sample of size n, (i=1, 2, …, n) Assume that all observations Xi are independent random variables The population mean = E[Xi]= The population standard deviation=(Var[Xi])0.5= Point estimate of the population mean= Mean and Std. Dev. of the point estimate for the population mean
378
Confidence Interval for Population Means (II)
Distribution of the point estimate for population means For any distribution of Xi (i=1, 2, …n), when n is large (n30), due to the Central Limit Theorem If all Xi (i=1, 2, …n) are normally distributed, for any n A standard transformation: Defining a symmetric two sided confidence interval P(Z/2 Z Z/2) = 1 is known Z/2 can be found from a N(0, 1) probability table Confidence interval for the population mean
379
Confidence Interval for Population Means (III)
In case the population standard deviation, , is known In case is unknown we need to estimate it Use the point estimate s The test variable Z is no longer Normally distributed, it follows a Students-t distribution with n-1 degrees of freedom In practice when n is large (30) the t-distribution is often approximated with the Normal distribution!
380
Determining an Appropriate Sample Size
A common problem in simulation How many runs and how long should they be? Depends on the variability of the sought output variables If a symmetric confidence interval of width 2d is desired for a mean performance measure If x-bar is normally distributed If is unknown and estimated with s
381
Hypothesis Testing (I)
Testing if a population mean () is equal to, larger than or smaller than a given value Suppose that in a sample of n observations the point estimate of = Hypothesis Reject H0 if … Type of test H0: =a Symmetric two tail test HA: a H0: a One tail test HA: <a H0: a HA: >a
382
Hypothesis Testing (II)
2. Testing if two sample means are significantly different Useful when comparing process designs A two tail test when 1=2=s H0: 1- 2=a /typically a=0/ HA: 1- 2a The test statistic Z belongs to a Student-t distribution Reject H0 on the significance level if it is not true that
383
Hypothesis Testing (III)
If the sample sizes are large (n1+n2-2>30) Z is approximately N(0, 1) distributed Reject H0 if it is not true that In practice, when comparing designs non-overlapping 3 intervals are often used as a criteria H0: 1- 2>0 HA: 1- 20 Reject H0 if
384
Optimizing Business Process Performance
Chapter 10 Business Process Modeling, Simulation and Design
385
Optimizing Business Processes
NOTE: We recommend presenting this chapter by running Extend 6.0 directly, and interactively show how the program works. However, for your convenience, we have attached a selection of the figures/screenshots from Chapter 8 of the book as the basis for an in class presentation without access to a computer with Extend installed.
386
Optimizing Business Processes
Conceptualization of an optimization model
387
Optimizing Business Processes
Conceptualization of an optimization model with uncertainty
388
Process Optimization Tutorial
0.10 Clerk & Technician Agent & Supervisor Technician Agent Agent C D H I J 0.60 Clerk Clerk A B Technician 0.40 0.85 E G Technician 0.15 F Technician & Supervisor Business process with multiple resource types
389
Optimization Tutorial
Extend model of process with an Evolutionary Optimizer block
390
Optimization Tutorial
Clone tool in the Extend toolbar
391
Optimization Tutorial
Resource Pool block dialog for Supervisors
392
Optimization Tutorial
Timer block dialog
393
Optimization Tutorial
Set Cost tab of the Evolutionary Optimizer block dialog
394
Optimization Tutorial
Global constraint to model limit in the total number of workers
395
Optimization Tutorial
Parameter settings for the Evolutionary Optimizer
396
Optimization Tutorial
Results tab showing ordered population
397
Optimization Tutorial
Optimization Value plot
398
Process Benchmarking with Data Envelopment Analysis
Chapter 11 Business Process Modeling, Simulation and Design
399
Overview DEA: Tool for Benchmarking
Relative Efficiency – Important Concepts Black-box model Graphical Analysis Efficiency calculations Linear Programming Formulation Using the Excel Add-in
400
DEA: Tool for Benchmarking
Successfully applied to assess the efficiency of various organizations and/or processes. Process = Decision Making Unit (DMU) The efficiency of a process is only relative to the performance of other processes in the set Considers process as a black box and analyzes the relationships between its inputs and outputs
401
Process as Black Box Figure Black box model of a process Output Efficiency = Input However, with multiple inputs and outputs, it becomes more difficult to evaluate the process efficiency.
402
Calculating Efficiency
Clearly, process A is more efficient than process B, but... A new assessment based on office space shows that process B is more efficient than process A, so…
403
Calculating Efficiency
DEA offers a variety of models that use multiple inputs and outputs to compare the efficiency of two or more processes. The ratio model is based on the following definition of efficiency: Weighted Sum of Outputs Efficiency = Weighted Sum of Inputs
404
Graphical Analysis Suppose we have the following input and output data: We label the independent efficiency ratios x and y:
405
Graphical Analysis Then, we plot the relative position of each process: Efficient frontier
406
Efficiency Calculations
Relatively efficient processes are those on the efficient frontier: Considered to have 100% efficiency. What is the efficiency of the relatively inefficient processes? P1 (x1,y1) (xv,yv) P2 (x2,y2) P0 (x0,y0) y = output2/input x = output1/input Projection of a relatively inefficient process
407
Efficiency Calculations
P1 and P2 are relatively efficient P0’s peer group. Define a and b such that: P1 (x1,y1) (xv,yv) P2 (x2,y2) P0 (x0,y0) y = output2/input x = output1/input Then, we get the efficient virtual process corresponding to xv and yv: The efficiency of process P0 is: Projection of a relatively inefficient process
408
Linear Programming The ratio model measures the efficiency of a process by comparing to a hypothetical process that is a weighted linear combination of other processes. Individual processes might value inputs and outputs differently. Therefore, each process is allowed to adopt a set of weights to show it in the most favorable light. Formulated as a sequence of linear programs (one for each process) to: Maximize the efficiency of one process Subject to the efficiency of all processes 100%
409
Linear Programming The variables are the weights assigned to each input and output: wout(j), win(i) An LP formulation for a given process p is:
410
Using the Excel Add-in NOTE:
We recommend presenting/explaining the DEA Add-in, available on the CD that comes with the textbook, by running it directly in Excel. However, for your convenience, we have attached a selection of the figures/screenshots from Chapter 11 of the book as the basis for an in class presentation without access to a computer with the Excel Add-in installed.
411
Using the Excel Add-in Creating a new DEA model
412
Using the Excel Add-in New model dialog window
413
Using the Excel Add-in Completed Example.Output worksheet
414
Using the Excel Add-in Efficiency worksheet
415
Using the Excel Add-in Best Practice worksheet
416
Using the Excel Add-in Targets worksheet
417
Using the Excel Add-in Virtual Output worksheet
418
Using the Excel Add-in Virtual Output chart
Similar presentations
© 2025 SlidePlayer.com Inc.
All rights reserved.