Multi-objectivization inspired metaheuristics for the sum-of-the-parts combinatorial optimization problems

Multi-objectivization is a term used to describe strategies developed for optimizing single-objective problems by multi-objective algorithms. This paper focuses on multi-objectivizing the sum-of-the-parts combinatorial optimization problems, which include traveling salesman problem, unconstrained binary quadratic programming and other well-known problem. For we propose decompose its original objective into two sub-objectives with controllable correlation. Based decomposition method, new multi-objectivization inspired techniques called non-dominance search exploitation are developed, respectively. Non-dominance combined metaheuristics, namely iterated local tabu search, while embedded within Lin–Kernighan metaheuristic. The resultant metaheuristics ILS+NDS, ITS+NDS ILK+NDE, Empirical studies some TSP UBQP instances show that appropriate correlation between sub-objectives, there more chances escape from optima when starting solution selected non-dominated solutions defined decomposed sub-objectives. Experimental results also ILK+NDE all significantly outperform their counterparts most of test instances.