Performance evaluation of simple multiobjective genetic local search algorithms on multiobjective 0/1 knapsack problems

The aim of this paper is to demonstrate high search ability of our simple multiobjective genetic local search (S-MOGLS) algorithm. First we explain the basic framework of the S-MOGLS algorithm, which can be implemented and efficiently executed with small memory storage and short CPU time. The S-MOGLS algorithm uses Pareto ranking and a crowding measure for generation update in the same manner as the NSGA-II. Thus the SMOGLS algorithm can be viewed as a hybrid algorithm of the NSGA-II with local search. Next we examine the performance of various variants of the S-MOGLS algorithm. Some variants use a weighted scalar fitness function in parent selection and local search while others use Pareto ranking. In computational experiments we examine a wide range of parameter specifications fro finding the point in the implementation of hybrid algorithms. Finally the S-MOGLS algorithm is compared with some evolutionary multiobjective optimization algorithms.