Helsinki University of Technology Systems Analysis Laboratory Incomplete Ordinal Information in Value Tree Analysis Antti Punkka and Ahti Salo Systems.

Slides:



Advertisements
Similar presentations
Teknillinen korkeakoulu Systeemianalyysin laboratorio 1 Graduate school seminar Rank-Based DEA-Efficiency Analysis Samuli Leppänen Systems.
Advertisements

Developing the Strategic Research Agenda (SRA) for the Forest-Based Sector Technology Platform (FTP) RPM-Analysis Ahti Salo, Totti Könnölä and Ville Brummer.
Multi‑Criteria Decision Making
1 Helsinki University of Technology Systems Analysis Laboratory Robust Portfolio Modeling for Scenario-Based Project Appraisal Juuso Liesiö, Pekka Mild.
1 Ratio-Based Efficiency Analysis Antti Punkka and Ahti Salo Systems Analysis Laboratory Aalto University School of Science P.O. Box 11100, Aalto.
1PRIME Decisions - An Interactive Tool for Value Tree Analysis Helsinki University of Technology Systems Analysis Laboratory PRIME Decisions - An Interactive.
Helsinki University of Technology Systems Analysis Laboratory RPM – Robust Portfolio Modeling for Project Selection Pekka Mild, Juuso Liesiö and Ahti Salo.
Helsinki University of Technology Systems Analysis Laboratory RICHER – A Method for Exploiting Incomplete Ordinal Information in Value Trees Antti Punkka.
1 Helsinki University of Technology Systems Analysis Laboratory Multi-Criteria Capital Budgeting with Incomplete Preference Information Pekka Mild, Juuso.
1 Mutli-Attribute Decision Making Eliciting Weights Scott Matthews Courses: /
From transaction cost to transactional value analysis: Implications for the study of inter- organizational strategies Zajac, Edward J. & Olsen, Cyrus P.
MADM Y. İlker TOPCU, Ph.D twitter.com/yitopcu.
S ystems Analysis Laboratory Helsinki University of Technology A Preference Programming Approach to Make the Even Swaps Method Even Easier Jyri Mustajoki.
S ystems Analysis Laboratory Helsinki University of Technology Decision Support for the Even Swaps Process with Preference Programming Jyri Mustajoki Raimo.
Decision analysis and Risk Management course in Kuopio
Helsinki University of Technology Systems Analysis Laboratory A Portfolio Model for the Allocation of Resources to Standardization Activities Antti Toppila,
S ystems Analysis Laboratory Helsinki University of Technology Using Intervals for Global Sensitivity and Worst Case Analyses in Multiattribute Value Trees.
Presented by Johanna Lind and Anna Schurba Facility Location Planning using the Analytic Hierarchy Process Specialisation Seminar „Facility Location Planning“
Modeling (Chap. 2) Modern Information Retrieval Spring 2000.
1 Helsinki University of Technology Systems Analysis Laboratory Robust Portfolio Selection in Multiattribute Capital Budgeting Pekka Mild and Ahti Salo.
1 S ystems Analysis Laboratory Helsinki University of Technology Decision and Negotiation Support in Multi-Stakeholder Development of Lake Regulation Policy.
ELearning / MCDA Systems Analysis Laboratory Helsinki University of Technology Introduction to Value Tree Analysis eLearning resources / MCDA team Director.
Helsinki University of Technology Systems Analysis Laboratory Ahti Salo and Antti Punkka Systems Analysis Laboratory Helsinki University of Technology.
Analyzing the Problem (MAVT) Y. İlker TOPCU, Ph.D twitter.com/yitopcu.
1 Helsinki University of Technology Systems Analysis Laboratory Robust Portfolio Modeling in the Development of National Research Priorities Ville Brummer.
An efficient distributed protocol for collective decision- making in combinatorial domains CMSS Feb , 2012 Minyi Li Intelligent Agent Technology.
STOCHASTIC DOMINANCE APPROACH TO PORTFOLIO OPTIMIZATION Nesrin Alptekin Anadolu University, TURKEY.
1 Helsinki University of Technology Systems Analysis Laboratory Rank-Based Sensitivity Analysis of Multiattribute Value Models Antti Punkka and Ahti Salo.
1 Helsinki University of Technology Systems Analysis Laboratory RPM-Explorer - A Web-based Tool for Interactive Portfolio Decision Analysis Erkka Jalonen.
Helsinki University of Technology Systems Analysis Laboratory Determining cost-effective portfolios of weapon systems Juuso Liesiö, Ahti Salo and Jussi.
S ystems Analysis Laboratory Helsinki University of Technology 1 Raimo P. Hämäläinen Systems Analysis Laboratory Helsinki University of Technology
1 Systems Analysis Laboratory Helsinki University of Technology How to Benefit from Decision Analysis in Environmental Life Cycle Assessment Pauli Miettinen.
1 Helsinki University of Technology Systems Analysis Laboratory INFORMS 2007 Seattle Efficiency and Sensitivity Analyses in the Evaluation of University.
Towards Robust Indexing for Ranked Queries Dong Xin, Chen Chen, Jiawei Han Department of Computer Science University of Illinois at Urbana-Champaign VLDB.
S ystems Analysis Laboratory Helsinki University of Technology Observations from computer- supported Even Swaps experiments using the Smart-Swaps software.
Helsinki University of Technology Systems Analysis Laboratory INFORMS Seattle 2007 Integrated Multi-Criteria Budgeting for Maintenance and Rehabilitation.
Tanja Magoč, François Modave, Xiaojing Wang, and Martine Ceberio Computer Science Department The University of Texas at El Paso.
A Dynamic Interval Goal Programming Approach to the Regulation of a Lake-River System Raimo P. Hämäläinen Juha Mäntysaari S ystems Analysis Laboratory.
S ystems Analysis Laboratory Helsinki University of Technology 1 Raimo P. Hämäläinen Jyri Mustajoki Systems Analysis Laboratory Helsinki University of.
1 Helsinki University of Technology Systems Analysis Laboratory Selecting Forest Sites for Voluntary Conservation in Finland Antti Punkka and Ahti Salo.
S ystems Analysis Laboratory Helsinki University of Technology Practical dominance and process support in the Even Swaps method Jyri Mustajoki Raimo P.
1 Helsinki University of Technology Systems Analysis Laboratory Selecting Forest Sites for Voluntary Conservation with Robust Portfolio Modeling Antti.
Helsinki University of Technology Systems Analysis Laboratory Antti Punkka and Ahti Salo Systems Analysis Laboratory Helsinki University of Technology.
Preference Modelling and Decision Support Roman Słowiński Poznań University of Technology, Poland  Roman Słowiński.
2nd Meeting of Young Researchers on MULTIPLE CRITERIA DECISION AIDING Iryna Yevseyeva Niilo Mäki Instituutti University of Jyväskylä, Finland
Helsinki University of Technology Systems Analysis Laboratory 1DAS workshop Ahti A. Salo and Raimo P. Hämäläinen Systems Analysis Laboratory Helsinki.
Helsinki University of Technology Systems Analysis Laboratory Portfolio and Scenario Analysis in the Cost-Effectiveness Evaluation of Weapon Systems Jussi.
1 Helsinki University of Technology Systems Analysis Laboratory Fostering the Diversity of Innovation Activities through e-Participation Totti Könnölä,
11 Ahti Salo, Juuso Liesiö and Eeva Vilkkumaa Department of Mathematics and Systems Analysis Aalto University School of Science and Technology P.O. Box.
S ystems Analysis Laboratory Helsinki University of Technology 1 Decision Analysis Raimo P. Hämäläinen Systems Analysis Laboratory Helsinki University.
1 School of Science and Technology Systems Analysis Laboratory Graduate school seminar presentation Current research topics in Portfolio Decision.
1 S ystems Analysis Laboratory Helsinki University of Technology Master’s Thesis Antti Punkka “ Uses of Ordinal Preference Information in Interactive Decision.
1 Ratio-Based Efficiency Analysis (REA) Antti Punkka and Ahti Salo Systems Analysis Laboratory Aalto University School of Science and Technology P.O. Box.
S ystems Analysis Laboratory Helsinki University of Technology 15th MCDM conference - Ankara Mats Lindstedt / 1 Using Intervals for Global.
Preference Modelling and Decision Support Roman Słowiński Poznań University of Technology, Poland  Roman Słowiński.
Helsinki University of Technology Systems Analysis Laboratory EURO 2009, Bonn Supporting Infrastructure Maintenance Project Selection with Robust Portfolio.
1 Helsinki University of Technology Systems Analysis Laboratory Standardization Portfolio Management for a Global Telecom Company Ville Brummer Systems.
To the Decision Deck platform UTA GMS /GRIP plugin Piotr Zielniewicz Poznan University of Technology, Poland 2nd Decision Deck Workshop February 21-22,
ON ELICITATION TECHNIQUES OF NEAR-CONSISTENT PAIRWISE COMPARISON MATRICES József Temesi Department of Operations Research Corvinus University of Budapest,
Mustajoki, Hämäläinen and Salo Decision support by interval SMART/SWING / 1 S ystems Analysis Laboratory Helsinki University of Technology Decision support.
preference statements
Mikko Harju*, Juuso Liesiö**, Kai Virtanen*
Flexible and Interactive Tradeoff Elicitation Procedure
Incomplete ordinal information in value tree analysis and comparison of DMU’s efficiency ratios with incomplete information Antti Punkka supervisor Prof.
D E C I S I O N A R I U M g l o b a l s p a c e f o r d e c i s i o n s u p p o r t group decision making multicriteria decision analysis group.
Raimo P. Hämäläinen Systems Analysis Laboratory
Decision support by interval SMART/SWING Methods to incorporate uncertainty into multiattribute analysis Ahti Salo Jyri Mustajoki Raimo P. Hämäläinen.
Juuso Liesiö, Pekka Mild and Ahti Salo Systems Analysis Laboratory
Introduction to Value Tree Analysis
FITradeoff Method (Flexible and Interactive Tradeoff)
Presentation transcript:

Helsinki University of Technology Systems Analysis Laboratory Incomplete Ordinal Information in Value Tree Analysis Antti Punkka and Ahti Salo Systems Analysis Laboratory Helsinki University of Technology P.O. Box 1100, TKK, Finland

Helsinki University of Technology Systems Analysis Laboratory 2 n n m alternatives X={x 1,…,x m } n n n twig-level attributes A={a 1,…,a n } n n Additive value n n Set of possible non-normalized scores – –v i (x i 0 )=0 and v i (x i * ) = w i Value tree analysis non-normalized form normalized form

Helsinki University of Technology Systems Analysis Laboratory 3 Preference information n Complete information –Point estimates for weights and scores –Examples »SWING (von Winterfeldt and Edwards 1986) »SMART (Edwards 1977) n Incomplete information –Modeled through linear constraints on weights and scores –Provides dominance relations and value intervals for alternatives –Supports ex ante sensitivity analysis in view of feasible parameters –Examples »PAIRS (Salo and Hämäläinen 1992) »PRIME (Salo and Hämäläinen 2001)

Helsinki University of Technology Systems Analysis Laboratory 4 Incomplete weight information n Forms of incomplete information (Park and Kim 1997): 1.weak ranking w i  w j 2.strict ranking w i – w j   3.ranking with multiples w i   w j 4.interval form  ≤ w i ≤  +  5.ranking of differences w i – w j  w k – w l n A feasible region for attribute weights S w

Helsinki University of Technology Systems Analysis Laboratory 5 Preference Programming with ordinal information Incomplete ordinal information about: 1. 1.Relative importance of attributes 2. 2.Values of alternatives w.r.t. - A single twig-level attribute - Several attributes (e.g. higher-level attributes) - All attributes (holistic comparisons) Dominance relations, Decision rules and Overall value intervals MILP model Decision Other forms of incomplete information: 1. 1.Weak ranking 2. 2.Strict ranking 3. 3.Ranking with multiples 4. 4.Interval form 5. 5.Ranking of differences Results sufficiently conclusive for the DM? Additional information yes no

Helsinki University of Technology Systems Analysis Laboratory 6 Ordinal preference information n Comparative statements between objects –No information about how much ’better’ or more important an object is than another –Can be useful in evaluation w.r.t. qualitative attributes –Complete information = a rank-ordering of all attributes or alternatives n Uses in preference elicitation –Rank attributes in terms of relative importance »Obtain point estimates through, e.g., rank sum weights (Stillwell et al. 1981), rank order centroid (Barron 1992) –Rank alternatives with regard to one or several attributes –Holistic comparisons: ”alternative x 1 preferred to alternative x 2 overall”

Helsinki University of Technology Systems Analysis Laboratory 7 Incomplete ordinal preference information n A complete rank-ordering, too, may be difficult to obtain –Identification of best performing alternative with regard to some attribute »which office facility has the best public transport connections? –Comparison of attributes »which attribute is the most important one? n Rank Inclusion in Criteria Hierarchies (RICH) –Salo and Punkka (2005), European Journal of Operational Research 163/2, pp –Admits incomplete ordinal information about the importance of attributes »”the most important attribute is either cost or durability” »”environmental factors is among the three most important attributes”

Helsinki University of Technology Systems Analysis Laboratory 8 Non-convex feasible region RICH n ”Either attribute a 1 or a 2 is the most important of the three attributes” n Four rank-orderings compatible with this statement Supported by RICH Decisions ©, n Selection of risk management methods (Ojanen et al. 2005) n Participatory priority-setting for a Scandinavian research program (Salo and Liesiö 2006)

Helsinki University of Technology Systems Analysis Laboratory 9 RICHER - RICH with Extended Rankings n Admits incomplete ordinal information about alternatives –”Alternatives x 1, x 2 and x 3 are the three most preferred ones with regard to environmental factors” –”Alternative x 1 is the least preferred among x 1, x 2 and x 3 w.r.t. cost” –”Alternative x 1 is not among the three most preferred ones overall” n Ordinal statements w.r.t. different attribute sets –Twig-level attributes –Higher-level attributes A’  A –Holistic statements w.r.t. all attributes

Helsinki University of Technology Systems Analysis Laboratory 10 Modeling of incomplete ordinal information (1/3) n The smaller the ranking, the more preferred the alternative –r(x 4 )=1  the ranking of x 4 is 1  it is the most preferred n Rank-orderings r=(r 1,..., r m’ ) on alternatives X’  X –Bijections from alternatives X’  X to corresponding rankings 1,...,|X’|=m’ –Notation: r i = r(x j ), s.t. j is the i-th smallest index in X’ –Convex feasible region »A’={a i }

Helsinki University of Technology Systems Analysis Laboratory 11 Modeling of incomplete ordinal information (2/3) n Specified as a set of alternatives I  X ’  X and corresponding rankings J  {1,...,m’} –X’ = subset of alternatives under comparison and m’ = |X’| its cardinality n If |I|<|J|, alternatives in I have their rankings in J –x 4 and x 5 belong to the three most preferred alternatives –I = {x 4, x 5 }, J = {1,2,3} n If |I|  |J|, rankings in J are attained by alternatives in I –The least preferred alternative in X={x 1,...,x 10 } is among x 1, x 2, x 3, x 4 –I = {x 1, x 2, x 3, x 4 }, J = {10}

Helsinki University of Technology Systems Analysis Laboratory 12 Modeling of incomplete ordinal information (3/3) n Sets I and J lead to compatible rank-orderings R(I,J) for each combination of X’, A’ n Feasible region associated with compatible rank-orderings n Sets S(I,J) have several useful properties, for example –S(I,J) = S(I C,J C ), where I C is the complement of I in X’ –Set inclusions: I 2  I 1, |I i |  |J| => S(I 2,J)  S(I 1,J)

Helsinki University of Technology Systems Analysis Laboratory 13 Linear inequality formulation for S(I,J) (1/3) n Values of alternatives with rankings k and k+1 are separated by milestone variable z k –If the ranking of x j is ”worse” than k, its value is at most z k –Binary variable y k (x j )=1 iff the value of x j is at least z k –Milestone, binary and value variables subjected to A’ and X’

Helsinki University of Technology Systems Analysis Laboratory 14 n n There are exactly k alternatives whose ranking is k or better n n If the ranking of x j is better than k-1, it is also better than k Linear inequality formulation for S(I,J) (2/3) decreasing value

Helsinki University of Technology Systems Analysis Laboratory 15 Linear inequality formulation for S(I,J) (3/3) n Feasible region S(I,J) characterized by linear constraints on binary variables n By using milestone and binary variables for each set pair (A’, X’) used in elicitation, all constraints are in the same linear model n Characteristics of incomplete ordinal information used to enhance computational properties –E.g., only the relevant milestone and binary variables are introduced »given a statement that alternatives x 1 and x 2 are the two most preferred, only z 2 is needed

Helsinki University of Technology Systems Analysis Laboratory 16 Pairwise dominance n Value intervals may overlap, but n Example with two attributes –Interval statement on weights –Point estimates for scores –x1 dominates x2 –x3 is also non-dominated n Non-dominated alternatives –Calculation through LP V w1w w2w x 1 dominates x 2 and strictly positive with some feasible scores and weights Alternative x k dominates x j Value intervals

Helsinki University of Technology Systems Analysis Laboratory 17 Decision rules n Maximize max overall value (’maximax’) => x 1 n Maximize min overall value (’maximin’) => x 3 n Maximize avg of max and min values (’central values’) => x 1 n Minimize greatest possible loss relative to another alternative (’minimax regret’) => x 1 V w1w w2w maximax maximin central values minimax regret

Helsinki University of Technology Systems Analysis Laboratory 18 RICHER n Key features –Extends preference elicitation techniques by admitting incomplete ordinal information about attributes and alternatives –Converts preference statements into a linear inequality formulation »can thus be combined with any other Preference Programming methods –Offers recommendations through pairwise dominance and decision rules n Decision support tools –Experiments suggest that MILP model is reasonably efficient –Software implementation of RICHER Decisions© ongoing n Future research directions –Sorting / classification procedures in score elicitation –Analyses of voting behavior (e.g., acceptance voting) Submitted manuscript downloadable at

Helsinki University of Technology Systems Analysis Laboratory 19 References Barron, F. H., “Selecting a Best Multiattribute Alternative with Partial Information about Attribute Weights”, Acta Psychologica 80 (1992) Edwards, W., “How to Use Multiattribute Utility Measurement for Social Decision Making”, IEEE Transactions on Systems, Man, and Cybernetics 7 (1977) Ojanen, O., Makkonen, S. and Salo, A., “A Multi-Criteria Framework for the Selection of Risk Analysis Methods at Energy Utilities”, International Journal of Risk Assessment and Management 5 (2005) Park, K. S. and Kim, S. H., “Tools dor Interactive Decision Making with Incompletely Identified Information”, European Journal of Operational Research 98 (1997) Salo, A. and Hämäläinen, R. P., "Preference Assessment by Imprecise Ratio Statements”, Operations Research 40 (1992) Salo, A. and Hämäläinen, R. P., “Preference Ratios in Multiattribute Evaluation (PRIME) - Elicitation and Decision Procedures under Incomplete Information”, IEEE Transactions on Systems, Man, and Cybernetics 31 (2001) Salo, A. and Liesiö, J., “A Case Study in Participatory Priority-Setting for a Scandinavian Research Program”, International Journal of Information Technology and Decision Making (to appear). Salo, A. and Punkka, A., “Rank Inclusion in Criteria Hierarchies”, European Journal of Operations Research 163 (2005) Stillwell, W. G., Seaver, D. A. and Edwards, W., “A Comparison of Weight Approximation Techniques in Multiattribute Utility Decision Making”, Organizational Behavior and Human Performance 28 (1981) von Winterfeldt, D., Edwards, W., ”Decision Analysis and Behavioral Research”, Cambridge University Press (1986).