Abstract: For solving the problems of standard Harmony Search（HS） algorithm convergences instability and can’t be used for multi-objective optimization，traditional HS is improved by introducing crossover operator，adaptive harmony memory considering rate and pitch adjusting rate.According to Pareto dominance and combining the harmony memory size characteristics of completely share information，a novel Improved Multi-objective Optimization Harmony Search algorithm（IMOHS） is proposed based on dynamic Pareto optimal front set.Simulation tests of several typical functions show that the proposed algorithm can efficiently and steadily converge to Pareto optimal front and find uniformly distributed Pareto set．

摘要： 针对标准和声搜索算法存在收敛不稳定及不能用于多目标优化问题的缺陷，通过引入交叉算子、自适应记忆内搜索概率和调节概率，改进了传统的和声搜索算法；根据Pareto支配关系，结合算法和声记忆库内信息完全共享的特性，提出了基于动态Pareto最优前沿的能够求解多目标优化问题的多目标改进和声搜索算法。通过几个典型函数的仿真测试表明，提出的算法能够高效稳定地收敛于Pareto最优前沿，获得分布均匀的Pareto解集。