An interactive procedure for multiobjective optimization
Inès Ben Jaâfar, Khaled Ghédira · International Conference on Systems · 2005
The majority of approaches for finding best compromise solutions to multiobjective optimization problems (MOP) make use of the Pareto optimality concept. However, in modeling real world problems, we often encounter MOP with large Pareto optimal alternatives to choose from. This paper introduces the concept of a-efficiency, which provides a notion that is stronger than Pareto optimality and allows setting up a preference ordering amongst various alternatives that are Pareto optimal. If the user still has to process quite a large number of alternatives, we propose to arrange them using an interactive approach based on a plurality voting procedure. This interactive procedure is based on a binary preference relation to rank the solutions set. However, for more flexibility this interactive procedure is extended exploiting fuzzy preference relation.