New multicriteria optimization method based on the use of a diploid genetic algorithm: Example of an industrial problem

Abstract

The design of a new product or of its manufacturing process consists in reconciling multiple objectives with each other to take into account their different features. In this paper, a new multicriteria optimization algorithm is presented. This method is based on the use of (i) a genetic algorithm (GA) which optimizes each system response and (ii) a selection algorithm which sorts Pareto-efficient points. This technique presents the great advantage of being of wide use. There is no particular mathematical condition about functions that are simultaneously optimized and, unlike the other multicriteria optimization methods which depend on the user's choice, our algorithm permits to obtain an optimal surface in which the user will be able to pick up his own working conditions. Efficiency of this new method is here illustrated with one mathematical example and with an industrial application.