An Experimental Study of Similarity-Based Mating on a Weight-Based Multiobjective Immune Genetic Algorithm

In this study, a novel mate selection approach is presented, which is a modification of the tournament selection algorithm with similar individuals (TSASI) used in the weight-based multiobjective immune genetic algorithm (WBMOIGA). Then we examine the effect of the proposed mate selection approach on the performance of WBMOIGA by comparing WBMOIGA with the proposed mate selection approach with WBMOIGA with TSASI and WBMOIGA with the binary tournament selection approach. Experimental results on three standard test problems (ZDT6, ZDT1, and FON) show that the performance of WBMOIGA can be improved by the proposed mate selection approach, and WBMOIGA with the proposed mate selection approach outperforms WBMOIGA with TSASI and WBMOIGA with the binary tournament selection approach.