Effect of using partial solutions in edge histogram sampling algorithms with different local searches

In previous study, we have proposed EHBSA within the EDA framework for permutation domains, and showed better performance than traditional GAs. The important feature of EHBSA is to use partial solutions from previous generations. In this paper, we analyze the effectiveness of using partial solutions using a wide range of problem sizes, without local search, and incorporating two types of local search. One of the most important finding in this paper is that we were able to confirm that using partial solutions is effective for all cases in which we use no local search, 3-OPT local search, and Lin-Kernighan (LK) local search. Future work for this research is also discussed.