Course

Multi-Objective Decision Making - 3 ECTS

Organised by Wageningen School of Social Sciences (WASS), Operations Research and Logistics
Date

Mon 28 October 2024 until Thu 14 November 2024

The question “what’s best?” often doesn’t have a single answer, and multiple objectives play a role simultaneously. Think of maximising process efficiency in engineering, while minimising environmental impact and costs, or deciding which foods can fit in a diet that is healthy, sustainable, affordable and appealing. Problems like these can be translated into a multi-objective optimization model. Such models help to explore the different answers to the question “what’s best?” and determine optimal solutions to the formulated problem. This course introduces you to a multi-objective decision making toolbox, to use in your own research.

Learning outcomes and activities

The objective of this PhD-course is to introduce the Multi-Objective Decision Making (MODM) toolbox to PhD students and scientists whose research contribute in developing sustainable and nutrition-secure food systems. Participants learn the theory behind commonly used MODM methods while emphasis is put on applying the methods consistently in their own research. After successful completion participants are expected to be able to:

  • Justify choice of appropriate MODM methods for addressing own research questions.
  • Formulate MODM mathematical programming models.
  • Develop programs to solve MODM models.

The learning outcomes of this course are achieved through a variety of learning activities including interactive lectures, computer practicals, group work and self-study. The lectures aim to introduce the state-of-art in MODM methodology and deliver the theory of commonly used methods in relevant to the course research topics. Computer practicals aim to train participants on how to apply MODM methods and develop technical skills that are required to deal with the case study assignments. During computer practicals students work in groups to enhance networking and peer-learning. During the case studies participants have the opportunity to apply MODM techniques
to real-life decision problems relevant to their own research.

Topics

The course comprise of four main components:

  1. Calculating Pareto efficient solutions and trade-offs between conflicting indicators.
  2. Identifying preferable solutions using multi-objective optimization techniques.
  3. Weight elicitation.
  4. Case study assignments.

    Planning

    This course spans a period of 3 weeks. The first week is on-campus, while the second and third week involve self study and online consulting moments. Preparatory material will be provided for participants to refresh their knowledge on the basic concepts and techniques of mathematical programming needed for this course.

    During the first day of the course we aim to:

    • introduce the structure of the course and align expectations.
    • discuss potential questions on basic concepts of mathematical programming that have been provided to participants as preparation material.
    • introduce multi-objective optimization problems.
    • getting familiar with the software.

    The second day we focus on calculating Pareto efficient solutions and trade-offs between conflicting indicators. More specifically:

    • the weighed sum method is introduced.
    • the ϵ-constraint method is introduced.
    • in the computer practical the participants have the opportunity to use both methods and understand advantages and limitations.

    Calculating Pareto efficient solutions results in a set of solutions from which rational decision makers can negotiate and make a choice.

    The third day aims to introduce Multi-Objective Optimization (MOO) techniques that are used to identify in a structured way the efficient solution which is most preferable to the decision maker. We introduce:

    • Weighted Goal Linear Programming (we also discuss scaling issues).
    • Compromise Programming.
    • Similarly to day two, in the computer practical the participants have the opportunity to use both methods and practice the impact of scaling approaches.

    To facilitate the process of identifying the most preferable solution to the decision maker(s) we use weights of importance (e.g. in WGLP, or CP) for the different objective functions (i.e. indicators to be optimized). The fourth day aims to discuss weight elicitation techniques:

    • An overview of weight elicitation methods is provided (e.g. interactive vs non-interactive and compositional vs non compositional methods).
    • To demonstrate the difference we focus on two (to be discussed) contrasting approaches.
    • The computer practical aims to show that the selected weight elicitation method can have consequences for the result of the analysis.

    During the fifth day applications of MODM methodology are presented by experts of the Operations Research and Logistics group of WUR. Three studies will be presented from the perspective of the MODM methods used (to be announced)

    During the afternoon of day 5 the case study assignment is introduced. During the second and third week students work on the assignments. Online consulting moments are scheduled. Submitting a final report of the assignment is a prerequisite for successfully finalizing the course.

      Session Morning Afternoon Remarks
      28 Oct, 9.00-17.00 Introduction and Multi-Objective optimization Computer practical
      29 Oct, 9.00-17.00 Calculating Pareto efficient solutions Computer practical
      30 Oct, 9.00-17.00 Weighted Goal Linear Programming and Compromize Programming Computer practical
      31 Oct, 9.00-17.00 Weight elicitation methods Computer practical
      1 Nov, 9.00-17.00 Applications of Multi-Objective Optimization studies in Agrifood Supply Chains and in Diet Modelling Introduction case study assignment
      Course week 2&3:
      4-8 Nov, 9.00-17.00 Work on case study assignment Work on case study assignment 6 Nov: online consultation moment
      11-14 Nov, 9.00-17.00 Work on case study assignment Work on case study assignment 11 Nov: online consultation moment; 14 Nov, submit case study assignment

      Assessment

      The assessment consist of 1) an evaluation of the participant’s participation in the sessions and 2) the final report of the case study assignment.

      Assumed prior knowledge

      An understanding of the basic concepts and techniques of mathematical programming is assumed. Self-study material will be shared for course participants to prepare for the course. Please contact the course coordinators in case of doubt about the required entry level.

      Course fees

      WGS PhDs with TSP 300 euro
      a) All other PhD candidates b) Postdocs and staff of the above mentioned Graduate Schools 640 euro
      All others 900 euro

      Cancellation conditions

      Participants can cancel their registration free of charge 1 month before the course starts. A cancellation fee of 100% applies if a participant cancels his/her registration less than 1 month prior to the start of the course.

      The organisers have the right to cancel the course no later than one month before the planned course start date in the case that the number of registrations does not reach the minimum.

      The participants will be notified of any changes at their e-mail addresses.