MODEL FOR DEALING WITH DUAL-ROLE FACTORS IN DEA: EXTENSIONS GONGBING BI,JINGJING DING,LIANG LIANG,JIE WU Presenter : Gongbing Bi School of Management University.

Slides:



Advertisements
Similar presentations
Completeness and Expressiveness
Advertisements

Some important properties Lectures of Prof. Doron Peled, Bar Ilan University.
Two-stage Data Envelopment Analysis
Introduction to Proofs
Incremental Linear Programming Linear programming involves finding a solution to the constraints, one that maximizes the given linear function of variables.
Price Of Anarchy: Routing
Elif Kongar*, Mahesh Baral and Tarek Sobh *Departments of Technology Management and Mechanical Engineering University of Bridgeport, Bridgeport, CT, U.S.A.
Fast Algorithms For Hierarchical Range Histogram Constructions
Online Scheduling with Known Arrival Times Nicholas G Hall (Ohio State University) Marc E Posner (Ohio State University) Chris N Potts (University of Southampton)
Outline. Theorem For the two processor network, Bit C(Leader) = Bit C(MaxF) = 2[log 2 ((M + 2)/3.5)] and Bit C t (Leader) = Bit C t (MaxF) = 2[log 2 ((M.
Parallel Scheduling of Complex DAGs under Uncertainty Grzegorz Malewicz.
II. Linear Block Codes. © Tallal Elshabrawy 2 Last Lecture H Matrix and Calculation of d min Error Detection Capability Error Correction Capability Error.
Decision Making: An Introduction 1. 2 Decision Making Decision Making is a process of choosing among two or more alternative courses of action for the.
1 Statistical Tests of Returns to Scale Using DEA Rajiv D. Banker Hsihui Chang Shih-Chi Chang.
Measuring Risk Management Performance of Insurers: a DEA Approach Yayuan Ren Illinois State University August, 2007.
1 Introduction to Computability Theory Lecture12: Reductions Prof. Amos Israeli.
The Rate of Convergence of AdaBoost Indraneel Mukherjee Cynthia Rudin Rob Schapire.
A Kolmogorov Complexity Approach for Measuring Attack Path Complexity By Nwokedi C. Idika & Bharat Bhargava Presented by Bharat Bhargava.
1 Undecidability Andreas Klappenecker [based on slides by Prof. Welch]
CPSC 668Set 10: Consensus with Byzantine Failures1 CPSC 668 Distributed Algorithms and Systems Fall 2006 Prof. Jennifer Welch.
A general approximation technique for constrained forest problems Michael X. Goemans & David P. Williamson Presented by: Yonatan Elhanani & Yuval Cohen.
So far we have learned about:
Gene Regulatory Networks - the Boolean Approach Andrey Zhdanov Based on the papers by Tatsuya Akutsu et al and others.
1 A Lyapunov Approach to Frequency Analysis Tingshu Hu, Andy Teel UC Santa Barbara Zongli Lin University of Virginia.
A Scalable Network Resource Allocation Mechanism With Bounded Efficiency Loss IEEE Journal on Selected Areas in Communications, 2006 Johari, R., Tsitsiklis,
1 On statistical models of cluster stability Z. Volkovich a, b, Z. Barzily a, L. Morozensky a a. Software Engineering Department, ORT Braude College of.
CS Dept, City Univ.1 The Complexity of Connectivity in Wireless Networks Presented by LUO Hongbo.
Decision Theory CHOICE (Social Choice) Professor : Dr. Liang Student : Kenwa Chu.
1 IEEE Trans. on Smart Grid, 3(1), pp , Optimal Power Allocation Under Communication Network Externalities --M.G. Kallitsis, G. Michailidis.
APPLYING EPSILON-DIFFERENTIAL PRIVATE QUERY LOG RELEASING SCHEME TO DOCUMENT RETRIEVAL Sicong Zhang, Hui Yang, Lisa Singh Georgetown University August.
A Unified Modeling Framework for Distributed Resource Allocation of General Fork and Join Processing Networks in ACM SIGMETRICS
Introduction to Proofs
A Taxonomy of Evaluation Approaches in Software Engineering A. Chatzigeorgiou, T. Chaikalis, G. Paschalidou, N. Vesyropoulos, C. K. Georgiadis, E. Stiakakis.
Outcome Based Evaluation for Digital Library Projects and Services
An evaluation of European airlines’ operational performance.
Desheng Dash Wu University of Toronto, Reykjavik University [with John R. Birge, Booth School of Business, University of Chicago] Accepted and to appear.
Trust-Aware Optimal Crowdsourcing With Budget Constraint Xiangyang Liu 1, He He 2, and John S. Baras 1 1 Institute for Systems Research and Department.
Keeping up with the Joneses, reference dependence, and equilibrium indeterminacy FUR XII conference, LUISS, Roma, 23 June 2006 Livio Stracca European Central.
1 2. Independence and Bernoulli Trials Independence: Events A and B are independent if It is easy to show that A, B independent implies are all independent.
Efficiency Analysis of a Multisectoral Economic System Efficiency Analysis of a Multisectoral Economic System Mikulas Luptáčik University of Economics.
1 FORMULATION OF TECHNICAL, ECONOMIC AND ENVIRONMENTAL EFFICIENCY MEASURES THAT ARE CONSISTENT WITH THE MATERIALS BALANCE CONDITION by Tim COELLI Centre.
1 ECE-517 Reinforcement Learning in Artificial Intelligence Lecture 7: Finite Horizon MDPs, Dynamic Programming Dr. Itamar Arel College of Engineering.
Department Of Industrial Engineering Duality And Sensitivity Analysis presented by: Taha Ben Omar Supervisor: Prof. Dr. Sahand Daneshvar.
INTERNATIONAL CONFERENCE ON INFORMATION SCIENCE AND TECHNOLOGY, P.P , MARCH An ANFIS-based Dispatching Rule For Complex Fuzzy Job Shop Scheduling.
Advanced Topics in Propositional Logic Chapter 17 Language, Proof and Logic.
Copyright © Cengage Learning. All rights reserved. CHAPTER 11 ANALYSIS OF ALGORITHM EFFICIENCY ANALYSIS OF ALGORITHM EFFICIENCY.
Exploiting Context Analysis for Combining Multiple Entity Resolution Systems -Ramu Bandaru Zhaoqi Chen Dmitri V.kalashnikov Sharad Mehrotra.
1 DEA Based Approaches and Their Applications in Supply Chain Management Dr. Sri Talluri Professor of Supply Chain Management Presentation at the Helsinki.
Greedy is not Enough: An Efficient Batch Mode Active Learning Algorithm Chen, Yi-wen( 陳憶文 ) Graduate Institute of Computer Science & Information Engineering.
1 The Theory of NP-Completeness 2 Cook ’ s Theorem (1971) Prof. Cook Toronto U. Receiving Turing Award (1982) Discussing difficult problems: worst case.
Theory of Computation, Feodor F. Dragan, Kent State University 1 TheoryofComputation Spring, 2015 (Feodor F. Dragan) Department of Computer Science Kent.
O PTIMAL SERVICE TASK PARTITION AND DISTRIBUTION IN GRID SYSTEM WITH STAR TOPOLOGY G REGORY L EVITIN, Y UAN -S HUN D AI Adviser: Frank, Yeong-Sung Lin.
Data Structures and Algorithm Analysis Introduction Lecturer: Ligang Dong, egan Tel: , Office: SIEE Building.
Designing Games for Distributed Optimization Na Li and Jason R. Marden IEEE Journal of Selected Topics in Signal Processing, Vol. 7, No. 2, pp ,
Inequalities for Stochastic Linear Programming Problems By Albert Madansky Presented by Kevin Byrnes.
Tunable QoS-Aware Network Survivability Presenter : Yen Fen Kao Advisor : Yeong Sung Lin 2013 Proceedings IEEE INFOCOM.
Copyright © Cengage Learning. All rights reserved. CHAPTER 8 RELATIONS.
Young CS 331 D&A of Algo. NP-Completeness1 NP-Completeness Reference: Computers and Intractability: A Guide to the Theory of NP-Completeness by Garey and.
1 Ratio-Based Efficiency Analysis (REA) Antti Punkka and Ahti Salo Systems Analysis Laboratory Aalto University School of Science and Technology P.O. Box.
Whole Test Suite Generation. Abstract Not all bugs lead to program crashes, and not always is there a formal specification to check the correctness of.
Section 1.7. Definitions A theorem is a statement that can be shown to be true using: definitions other theorems axioms (statements which are given as.
Approximation Algorithms based on linear programming.
CS151: Mathematical Foundations of Computing Mathematical Induction.
Chapter 2 Sets and Functions.
Mathematical Foundations of AI
On statistical models of cluster stability Z. Volkovich a, b, Z
Handbook of Applied Cryptography - CH4, from 4.1~4.3
3 THE CCR MODEL AND PRODUCTION CORRESPONDENCE
NP-Completeness Reference: Computers and Intractability: A Guide to the Theory of NP-Completeness by Garey and Johnson, W.H. Freeman and Company, 1979.
Chapter 11: Further Topics in Algebra
Presentation transcript:

MODEL FOR DEALING WITH DUAL-ROLE FACTORS IN DEA: EXTENSIONS GONGBING BI,JINGJING DING,LIANG LIANG,JIE WU Presenter : Gongbing Bi School of Management University of Science and Technology of China

Contents Introduction Cook’s dual-role model and extension Extended dual-role DEA model Conclusions

Introduction(1/2) In traditional application of DEA, all factors involved can be clearly classified as inputs or outputs. However, Beasley (1990, 1995) first addressed the factors that could be treated as both inputs and outputs. Factors having the property of both input and output are referred to as dual-role factors.

Introduction(2/2) Examples for dual-role factors (1) The number of nurse trainees on staff (2) Graduate students (3) Research income (4) etc.

Introduction(3/3) Beasley’s papers (1990, 1995), the author proposed a new framework to deal with dual-role factors in DEA. However, Cook et al. (2006) showed that the Beasley’s methodology might cause unreasonable evaluation results. In addition, the inconsistency in the logic to deal with flexible factors in input and output sides was explored. To correct the apparent flaws, Cook et al. (2006) provided new ways to handle flexible factors in DEA. In their paper, flexible factor was treated as both input and output simultaneously just as how Beasley did. However, Cook et al. recommended treating the flexible factor as being nondiscretionary in the input side when input-oriented DEA model was used.

Cook’s dual-role model and modification(1/4) Cook’s dual-role model

Cook’s dual-role model and modification(2/4) Flaw in the method If the DMU under evaluation is CCR efficient, then the classification of dual- role factor will depend on optimizing algorithm that we use to solve the previous model

Cook’s dual-role model and modification(3/4) Property 1 if DMUo is characterized as efficient by CCR model when dual-role factors are excluded from consideration, DMUo is indifferent towards the classification of dual-role factor,in other words, the efficiency of DMUo doesn’t change under any classification of dual-role factors. This is why the algorithm dependent classification could happen

Cook’s dual-role model and modification(4/4) modification According to property 1, it is proposed that (1) DMUs are evaluated while dual-role factors are excluded by CCR model before cook’s model is applied. At this step, the DMUs which are indifferent towards the classification of dual-role factors are known. (2) the model is solved for each DMU that is not indifferent to the classification. (3) the number of DMUs which have optimal solution d=0 is counted and the majority rule is applied to determine the status of each dual-role factor

Extended DEA model(1/11) Suppose there are n DMUs in a reference set. Each DMU has s outputs, denote Y k =(y 1k,y 2k,…,y sk ) T, m inputs, denote X k =(x 1k,x 2k,…,x mk ) T and L dual-role factors, denote W k =(w 1k,w 2k,…,w Lk ) T.

Extended DEA model(2/11) If dual-role factors are classified as inputs, the production possibility set having constant returns-to-scale characteristic If dual-role factors are classified as outputs, the CRS production possibility set

Extended DEA model(3/11) Dual-role are considered as both inputs and outputs simultaneously, the production possibility set ( )

Extended DEA model(4/11) The definition of production probability set T is also based on the following facts which are consistent with the property of dual-role factors as both inputs and outputs simultaneously: i) If some dual-role factor of DMUo under evaluation is considered as relatively excess when compared with efficient frontier (such dual-role factor has input characteristic). This means that DMUo will gain an advantage if the dual-role factor is considered as output. ii) If some dual-role factor of DMUo under evaluation is considered as relatively in short when compared with efficient frontier (such dual-role factor has output characteristic). This means that DMUo will gain an advantage if the dual-role factor is considered as input.

Extended DEA model(5/11) Based on the production possibility set of dual- role factors, a model that deals with dual-role factors that act as both inputs and outputs is constructed as model (1).

Extended DEA model(6/11) We proceed to find the relation between model (1) and CCR model in which all factors have clear classification. This is formulated in terms of the following result: Theorem 1 if DMUo achieves a score by model (1), then there exist at least one classification of dual-role factors under which the CCR efficiency is 1.

Extended DEA model(7/11) Proof Case 1: when dual-role factors are excluded, DMUo is evaluated as efficienct by CCR model. According to Property 1, it is easy to know that Theorem 1 holds since the efficiency score of DMUo under any classification of dual-role factors is 1

Extended DEA model(8/11) Case 2: when dual-role factors are excluded (model (2)), DMUo is evaluated as inefficienct by CCR model. Lemma 1 All feasible solutions to model (2) with constitute a convex set C. Lemma 2 If we set,the CCR efficiency is less than 1, where I is the index set of dual-role factors that are classified as input, O is that classified as output, and O and indicate the partition of dual-role factors that satisfy the following condition:

Extended DEA model(9/11) where is any point belonging to C For each partition i.e. I and O are disjointed and the union gives D={1,…,L}, we define sub set of convex set C as follows where i takes on values in. Lemma 3 If DMUo is evaluated as inefficient by CCR model, then C i is nonempty for any. Lemma 4 If is nonempty for any, then there exists a point such that

Extended DEA model(10/11) Proof of Case 2: Now we prove Theorem 1 by contradiction. Assume that Theorem 1 were not true in Case 2. By Lemma 3, is nonempty for any. According to Lemma 4, there exists a point x that belongs to C such that.This contradicts the assumption that the optimal value to model (1) is 1. So Theorem 1 holds.

Extended DEA model(11/11) The converse of previous theroem need not be ture. Example Consider a set of DMUs shown in Table 1. Table 2 shows the results of all DMUs under different classification of dual-role factors and model (1). From the results of DMU4, we know that the unit is characterized as efficient if dual-role factor is classified as input. However, if the dual-role factor is treated as both input and output simultaneously the efficiency score by model (1) is 0.8.

Conclusions Beasley (1990, 1995) first addressed the factors that could be treated as both inputs and outputs. As pointed out by Cook et al. (2007), dual-role factors have become an important and very much under-researched topic. We first address the problem of uncertainty with Cook’s method. A property is introduced to provide the decision maker with more information on which DMU is in fact indifferent towards the classification of dual facts. An extended DEA model is proposed to incorporate dual-role factors in DEA literature when dual-role factors can be treated as inputs and outputs simultaneously. We study dual-role factor that serves as input and output simultaneously from production possibility set’s perspective, which is quite different from the existing methods. In addition, the relation between the proposed model with the traditional CCR model are explored in this paper. Dual-role factors, by its nature, have the property that for any DMUo evaluated the values of them should be equal to the counterparts of the efficient DMU that serves as benchmarking. One may think how we can improve the amount of dual-role factors aiming at improving the overall efficiency. However, we believe that properties such as returns-to-scale and congestion of DMU should be determined by some extended models incorporating the property of dual-role factors and the efficiency of each DMU concerning dual-role factors then can be studied from those perspectives. This is a direction for future research.

Thank you