Evolutionary Algorithm for Bin Packing Problem

In this paper we investigate the use of evolution strategy (ES) for the bin packing problem. We used (1, λ)-ES, where 10 children are generated from one parent by means of a simple genetic operators; the crossover is not applied. The best of the descendants becomes the new parent solution. A widely known ordering representation modified for the purposes of the grouping the objects was used in the algorithm. One of the term of the fitness function is the penalty. The researches shown a great usefulness of the evolution strategy for the bin packing problem. The results of the experiments confirmed once more the power of the evolutionary algorithms which consists in ability to generate very good solutions without going into the structure of the problem.