Kurssin etusivu

  • Kurssin etusivu

    Lecturer: Carlos Antunes


    Introduction

    In classical Operational Research, the decision maker’s preferences are modeled a priori; that is, all the values are aggregated in a single objective function (or criterion), aimed at determining the optimal solution to the problem. However, it is currently recognized that these approaches are too reductive, being inadequate to address many real-world problems. In these problems, multiple perspectives should be taken into account to evaluate the merits of potential solutions; i.e., the decision maker is generally interested not just in minimizing the cost but also in maximizing the system reliability, minimizing the environmental impacts, etc.

    Approaches that make an a priori aggregation of the multiple perspectives cannot duly capture the conflicting nature of the objective functions, which make operational evaluation aspects of distinct nature and impair the exploitation of trade-offs among them. Therefore, multiobjective optimization models, which include explicitly the multiple evaluation aspects as distinct objective functions, enable to adequately capture the essential characteristics of real-world problems and improve their perception by decision makers.


    Contents

    * Main concepts of multi-objective linear and integer programming;

    * Scalarization processes;

    * Interactive methods;

    * iMOLPe –interactive multi-objective linear programming explorer;

    * Applications of multi-objective models; and 

    * Methods in the energy sector.


    Practical Matters

    * Teaching: Lectures (9h) 

    * Assessment methods: Home assignment (15h)

    * Grading principles:

    30% of the grade will be due to class attending and active participation

    70% exercises list to be submitted 1 week after the last class

    * Grading scale: 0-5

    * Language of instruction: English


    References

    * (Main ref.) Multiobjective Linear and Integer Programming (2016). Springer Cham. Link here.

    * Alves MJ, Antunes CH, Clímaco J (2015) Interactive MOLP explorer: a graphical-based computational tool for teaching and decision support in multi-objective linear programming models. Comput Appl Eng Educ 23(2):314–326.

    * Steuer R (1986) Multiple criteria optimization: theory computation and application. Wiley.


    Software

    iMOLPe: freely downloadable at http://www.uc.pt/en/org/inescc/software


    Duration

    October 29, 30, 31. Mon-Wed 14:15 - 17:00


    Location

    Seminar room M134, in Otakaari 1.


    Enrollment

    The course is open to all interested participants. Enrollment for the course for Aalto students is through oodi. 



Materiaalit►