Session:

LBP - Late Breaking Papers

Title:

An Evolutionary Technique for Multicriterial Optimization Based on Endocrine Paradigm

Authors:

Corina Rotar

Abstract:

Many evolutionary algorithms have been lately developed for solving multiobjective problems, appealing or not to the Pareto optimality concept. Al-though, the evolutionary techniques for multiobjective optimization confront with several issues as: elitism, diversity of the population, or efficient settings for the specific parameters of the algorithm. In this paper, we propose a new evolutionary technique, which is inspired by the behavior of the endocrine sys-tem and uses the Pareto non-dominance concept. Therefore, the members of the population aren't called chromosomes anymore, but hormones and, even if they evolve according to the genetic principles, a supplementary mechanism based on the endocrine paradigm is connected with standard approach to deal with multiobjective optimization problems. Moreover, the proposed algorithm, in order to maintain the diversity of the population, uses a specific scheme of fit-ness sharing, eliminating the inconvenient of defining an appropriate value of sharing factor.