Improving performances of the genetic algorithm by caching

Abstract

In this paper we optimize run-time performance of the genetic algorithm by caching. We are caching the genetic algorithm procedure for evaluation of an objective function. Least Recently Used (LRU) caching strategy is used, that is simple but effective. This approach is good for problems that have a relatively small length of item string, and a large evaluation time of objective function. We present results of the caching to genetic algorithm for solving one such problem - the simple plant location problem (SPLP).