A COMPARISON OF METHODS FOR SELECTING PREFERRED SOLUTIONS IN MULTIOBJECTIVE DECISION MAKING

Enrico Zio (1) , R. Bazzo (2)

1 SSEC - Chaire Sciences des Systèmes et Défis Energétiques EDF/ECP/Supélec
2 Dipartimento di Energia [Milano]

In multiobjective optimization problems, the identified Pareto Frontiers and Sets often contain too many solutions, which make it difficult for the decision maker to select a preferred alternative. To facilitate the selection task, decision making support tools can be used in different instances of the multiobjective optimization search to introduce preferences on the objectives or to give a condensed representation of the solutions on the Pareto Frontier, so as to offer to the decision maker a manageable picture of the solution alternatives. This paper presents a comparison of some a priori and a posteriori decision making support methods, aimed at aiding the decision maker in the selection of the preferred solutions. The considered methods are compared with respect to their application to a case study concerning the optimization of the test intervals of the components of a safety system of a nuclear power plant. The engine for the multiobjective optimization search is based on genetic algorithms.

Domains: Neural and Evolutionary Computing [cs.NE]

Enrico Zio, R. Bazzo. A COMPARISON OF METHODS FOR SELECTING PREFERRED SOLUTIONS IN MULTIOBJECTIVE DECISION MAKING. Cengiz Kahraman. Computational Intelligence Systems in Industrial Engineering, Springer, pp.23-43, 2012, Atlantis Computational Intelligence Systems, Vol. 6. ⟨hal-00771555⟩