A simple multiobjective evolutionary algorithm for design of digital maps

Abstract

This paper presents a simple multiobjective evolutionary algorithm and considers its application to two-objective optimization problems in design of digital maps. The algorithm is based on the MOEA/D and uses mutation operators in reproduction of potential solutions. The digital map is defined on a set of points and can generate various periodic spike-trains. The algorithm is applied to two simple problems that require optimization of autocorrelation of periodic spike-trains and distribution of inter spike intervals. The first problem has a trade-off between two objective functions and the algorithm can find an approximated Pareto front. The second problem does not have a trade-off and the algorithm can find an approximated utopia point. The algorithm performance is confirmed in elementary numerical experiments.