Abstract: Harmony Search(HS) is a heuristic optimization method. To make HS algorithm be able to apply to the combinatorial optimization problem, a Discrete Harmony Search(DHS) algorithm is proposed. Two improvements of the DHS algorithm are described. The pure DHS algorithm may produce premature and poor results. In order to avoid premature convergence and poor results in DHS, based on the complementary strengths of DHS and Simulated Annealing(SA) algorithm, a hybrid strategy of HS and SA is put forward. Experimental results show that within the framework of the newly designed hybrid algorithm, the NP-hard classic Traveling Salesman Problem(TSP) can be efficiently solved with higher quality.

Key words: Traveling Salesman Problem(TSP), Harmony Search(HS), Simulated Annealing(SA), hybrid algorithm

摘要： 和声搜索是一种启发式优化算法，对和声搜索算法进行离散化，使其能够适用于组合优化问题，就离散和声搜索算法进行两点改进，针对离散和声搜索算法容易陷入局部最优的缺陷，提出一种离散和声搜索与模拟退火算法的混合策略。实验结果表明，基于改进离散和声搜索与模拟退火的混合算法具有较高的求解质量。

关键词: 旅行商问题, 和声搜索, 模拟退火, 混合算法

CLC Number: 

• TP301.6