Virtual workshop

Multiple Criteria Decision Analysis
Methods Selection Software (MCDA-MSS)

Event Info

icon_calendar.jpgJune 1, 2021
icon_clock.jpg8-11am EDT
2-5pm CEST
icon_stopwatch.jpgDuration: 3 hours

Presented by


Presented by Dr. Marco Cinelli in cooperation with Associate Professor Miłosz Kadziński and Professor Roman Słowiński

Marco_Cinelli_10_2017.png

Marco Cinelli (Ph.D. Engineering, University of Warwick, UK, 2016) is a Marie Skłodowska-Curie Global Fellow at (i) the Center for Environmental Solutions and Emergency Response (CESER) (formerly National Risk Management Research Laboratory (NRMRL)) of the U.S. Environmental Protection Agency in Cincinnati, USA, (ii) the Laboratory for Energy Systems Analysis, Paul Scherrer Institute, Villigen PSI, Switzerland and (iii) the Institute of Computing Science at Poznan University of Technology, Poland.

His research focuses on the development of Decision Support Systems (DSS) for materials, processes, energy systems and technologies assessment. The main frameworks he uses include sustainability, resilience and risk assessment, while the decision support methods belong to the area of Multiple Criteria Decision Aiding/Analysis (MCDA).
  

Marco_Cinelli_10_2017.png

Miłosz Kadziński is an Associate Professor at Poznan University of Technology and a Polish Young Academy member. He holds the MCDM Doctoral Dissertation Award 2013. He has been distinguished with the scientific awards by EURO, INFORMS MCDM Section, Polish Academy of Sciences, and Ministry of Science. He has been acknowledged with thirteen Best Reviewer Awards by EJOR, Omega, and GDN. He acts as a co-president of the Decision Deck Consortium. His research interests include preference learning, robustness analysis, and strengthening the interfaces between MCDM and other sub-disciplines of algorithmic decision theory, such as interactive and evolutionary multiple objective optimization, data envelopment analysis, and machine learning. He has published more than 65 papers in premier international journals such as Omega, EJOR, DSS, INS, COR, KBS, ESWA, and ML.
  

Roman Słowiński is a Professor and Founding Chair of the Laboratory of Intelligent Decision Support Systems at Poznań University of Technology, and a Professor in the Systems Research Institute of the Polish Academy of Sciences. As an ordinary member of the Polish Academy of Sciences he is its Vice President, elected for the term 2019-2022. He is a member of Academia Europaea and Fellow of IEEE, IRSS, INFORMS and IFIP. In his research, he combines Operational Research and Artificial Intelligence for Decision Aiding. Recipient of the EURO Gold Medal by the European Association of Operational Research Societies (1991), and Doctor HC of Polytechnic Faculty of Mons (Belgium, 2000), University Paris Dauphine (France, 2001), and Technical University of Crete (Greece, 2008). In 2005 he received the Annual Prize of the Foundation for Polish Science - the highest scientific honor awarded in Poland, and in 2020 - the Scientific Award of the Prime Minister of Poland. Since 1999, he is the principal editor of the European Journal of Operational Research (Elsevier), a premier journal in Operational Research.

Background

The majority of decision-making problems is characterized by a number of alternatives assessed with a set of criteria. In most of these problems, there are no dominating alternatives, meaning that no alternative performs at least as well as the others in all criteria and better for at least one of them. Consequently, the alternative(s) that can be recommended are those that represent the best compromise according to the selected criteria. Stakeholder preferences in the decision-making process can be included to consider the different priorities on the criteria and the aggregation of evaluation of performances on such criteria.

Multiple Criteria Decision Analysis (MCDA) methods are excellent tools to compare alternatives comprehensively (with e.g., a ranking, a classification) and lead to a decision recommendation. Their power resides in their capacity to convey a wealth of information representing each alternative in the process of for example ranking them from the best to the worst or in the process of their classification (e.g., into good, medium, and bad classes).

Knowledge gap

Over the last few decades, the number of MCDA methods has grown steadily (hundreds are available nowadays), and an analyst can find it difficult to select the relevant MCDA method(s) for the problem under consideration. The main issue that decision analysts have to deal with is summarized by the figure below and the question on its right.

“Which is the most suitable MCDA method (or subset of methods) that should be used for a given decision-making problem?”

The solution: MCDA-MSS

We have developed the first MCDA-Methods Selection Software (MCDA-MSS) that allows decision analysts to find the most relevant MCDA methods (among > 200 of them) for many decision-making problems, from relatively simple to very complex. This software has two aims:

  1. Allow analysts to learn our sequential and dynamic framework to describe complex decision-making;
  2. Guide an analyst assisting a Decision Maker (DM) in choosing the most appropriate MCDA method(s) for a given decision-making problem.

MCDA-MSS is a main outcome of the Marie Skłodowska-Curie Global Fellowship at Poznań University of Technology (Sustainability Assessment based on Decision Aiding, grant agreement No 743553, September 2018 – September 2021) of Dr. Marco Cinelli and his scientific exchanges at CESER at the U.S. EPA (September 2018 – July 2020) and Paul Scherrer Institute (September 2020 – February 2021).

Structure of the Workshop (see also Appendix I)

The virtual workshop on MCDA-MSS will last 3h (including breaks). This allows for the discussion of the input given by the attendees, as well as the presentation and use of MCDA-MSS with its four sections:

  1. Problem typology: Defines the type and structure of the decision-making problem
  2. Preference model: Defines the type of model that the user would like to apply
  3. Elicitation of preferences: Defines the type, modality and frequency of model preferences
  4. Exploitation of the preference model: Defines the strategy used to derive and enrich the decision recommendation

Format of the workshop

Before the workshop: material to be sent to the organizers by May 16, 2021 (only for those who will attend all the workshop, see Appendix 1)

The attendees can prepare their own background material with the following two Excel sheets:

  1. Sheet 1 (necessary) – Summary of features: Summary of objective features[1] that the analysts think should be used to select an MCDA method  for a certain decision-making problem. It is necessary to be explicit and fill in an Excel sheet where the attendees provide references & briefly describe each feature. If desired, the four sections of MCDA-MSS can be used as guidance, but this is not compulsory. The attendee can define his/her own sections (see Appendix 2 for details on the requested input and its boundaries).
     
  2. Sheet 2 (optional) – Use of the features: Examples of case studies developed by the MCDA analysts showing the “activated” features justifying the selection of the MCDA methods. The attendee can thus develop a second Excel sheet showing how the features listed in their first sheet were used to justify the selection of the MCDA methods.
     

During the workshop

The workshop will be held virtually, and it will include four sessions. Session 1 will provide an overview of MCDA-MSS and its sections, while the remaining three sessions will be interactive with MCDA-MSS training and hands-on use of the software to:

  1. Explore its (i) intelligibility, (ii) comprehensiveness, (iii) easiness of use, (iv) suitability for learning, and (v) interactive efficiency;
  2. Compare the features listed by the attendees with those reported in MCDA-MSS and study if they match and/or if any is missing in the lists of the attendees, as well as in MCDA-MSS.

Deliverables

The workshop will allow gathering the knowledge and expertise of the analysts who will attend the event. It is important to note that this workshop will be part of a series that the MCDA-MSS team will deliver at other organizations (i.e., U.S. EPA, Poznań University of Technology, Paul Scherrer Institute), as well as dissemination events (i.e., 31st European Conference on Operational Research (EURO) 2021).

The first deliverable will be a short workshop summary with the repository of decision-making problems that have been tackled by the attendees (and which will be shared with them too). This will allow understanding (i) the most frequent features they dealt with in their complex problems and (ii) the main difficulties encountered while developing the alternatives and the criteria set, eliciting stakeholders’ preferences, and choosing strategies to enhance the decision recommendation. A key benefit from this repository is a visual map of the different research avenues and projects that have been undertaken by the attendees.

The second deliverable will be a scientific paper summarizing the key learning insights from the event.

Interested to attend the workshop?

You can sign up here and choose between attending:

  • Session 1 only (presentation of MCDA-MSS): no limit to the number of attendees
  • Sessions 1-4 (the whole workshop): this option is limited to ~20 people to guarantee a manageable and constructive experience for all the attendees. So, sign up early! Please note that if you choose this option you agree to prepare by May 16, 2021 the background material as described in the workshop format section.

Software license

The software will be available free of charge to any interested user on a dedicated web platform.

The MCDA-MSS team

Poznań University of Technology (PL)
Dr. Marco Cinelli, Assoc. Prof. Miłosz Kadziński, Grzegorz Miebs, Prof. Roman Słowiński

U.S. EPA (USA)
Dr. Michael Gonzalez

Paul Scherrer Institute (CH)
Dr. Peter Burgherr

Appendices

Appendix I: Virtual MCDA-MSS workshop program

EDT CEST Virtual MCDA-MSS Workshop Program
8–8:05am 2–2:05pm Welcome
8:05–9am 2:05–3pm

Session 1: Introduction to MCDA-MSS

Description of MCDA-MSS & its sections:
Problem typology, preference model, elicitation of preferences, exploitation of the preference model

9-9:40am 3-3:40pm

Session 2: Co-Constructing MCDA methods selection

Discussion of features from the attendees & their case studies:
Missing features in the attendees’ lists; missing features in MCDA-MSS; case studies analysis

9:40-9:50am 3:40-3:50pm Coffee break
9:50-10:30am 3:50-4.30pm

Session 3A: Application of MCDA-MSS by the attendees

20 min guided exercise led by the moderator, 20 min use of MCDA-MSS individually or in small groups

10:30-10:35am 4:30-4:35pm Coffee break
10:35-10:55am 4:35-4:55pm

Session 3B: Application of MCDA-MSS by the attendees

20 min discussion with all the attendees to identify trends/discuss issues

10:55-11am 4:55-5pm

Session 4: Summary & Next Steps

Summary of the event & planning of deliverables

Appendix II: Guidance on preparation of the Excel sheet “Summary of features”

Given that a main objective of the workshop consists in testing the comprehensiveness of the features included in MCDA-MSS by comparing them with those selected by the MCDA analysts, it is important to point our boundary conditions that can help the analysts tailoring their efforts and set the expectations. Here is a list of the key boundary conditions for the development of the Excel sheet “Summary of features”, in line with the MCDA-MSS working strategy:

  1. MCDA methods are intended to support a comprehensive assessment of a finite set of alternatives using a family of criteria, while also accounting for the preferences of the decision maker;
  2. Presently, MCDA-MSS is handling MCDA methods for a single decision maker only;
  3. Decision-making problems are thought as unique descriptions of a certain decision-making challenge, including the problem statement (ranking, choice, classification), selection of the evaluation criteria, elicitation strategies for the preferences of the decision maker, and the expected exploitation of a preference model in view of presenting a decision recommendation;
  4. Multi-stage multicriteria decision-making problems are presently not included in MCDA-MSS.

Here is a more detailed overview of the sections of MCDA-MSS and the type of features they include:

  1. Section 1 (Problem typology): It defines how the problem is framed by (i) choosing the type of decision-making problem under consideration and (ii) defining the type and structure of the criteria used to evaluate the alternatives;
  2. Section 2 (Preference model): It defines the type of model the decision maker would like to apply, accounting for how the input data (i.e., criteria performances and preferences) is used by the method, as well as the aggregation procedure to derive the decision recommendation;
  3. Section 3 (Elicitation of preferences): It considers how the decision maker provides the preference information, its frequency and also the certainty with which such preferences are expressed;
  4. Section 4 (Exploitation of the preference model): It accounts for different strategies that can be used to exploit the preference model to derive a univocal decision recommendation and/or to show how variable the decision recommendation can be when there is uncertainty with respect to the performances of alternatives and/or representation of the decision maker’s preferences by the assumed preference model.

[1] Objective features are those which unequivocally determine the suitability (or not) of a method to a certain decision-making problem, independently from the knowledge and expertise of the analysts who lead the MCDA process, as well as qualitative requirements that the stakeholders might set (e.g., easiness of the methods, software availability).