Preference Incorporation into Evolutionary Multiobjective Optimization using Preference Information Implicit in a Set of Assignment Examples

Abstract

Nowadays, most approaches in the evolutionary multiobjective optimization literature concentrate mainly on adapting an evolutionary algorithm to generate an approximation of the Pareto frontier. However, this does not solve the problem. We present a new idea to incorporate into a MOEA the Decision Maker (DM) preferences, expressed in a set of solutions assigned to ordered categories. We modified the Non-dominated Sorting Genetic Algorithm 2 (NSGA2) to make selective pressure towards non-dominated solutions that belong to the most preferred category. In several instances, on the project portfolio problem, our proposal outperforms the standard NSGA2, achieving non-dominated solutions that best match the DM's preferences.