105 二次割当問題における交叉dMSXFの解探索性能の検証

抄録

dMSXF and dMSMF are promising methods in combinatorial problems, which perform deterministic multi-step searches using a neighborhood structure and a distance measure. dMSXF generates offspring in the interpolation domain focusing on inheritance of parents' characteristics. In contrast, dMSMF, which has been proposed as a complementary search of dMSXF searches in the extrapolation domain to acquire characteristics which are not appeared in the parents. In previous research, it became clear that dMSXF+dMSMF performed well on both TSP and JSP which have structural differences between their landscapes. In this paper, we applied dMSXF and dMSXF+dMSMF to QAP due to its structural simplicity. Through the experiments, dMSXF was shown to be high availability and incorporation of dMSMF into dMSXF improved their effective search performances on QAP instances.