【作者】 党明梅； 王振雷； 钱锋；

【Author】 Dang Mingmei Wang Zhenlei Qian Feng (East China University of Science and Technology,Key Laboratory of Advanced Control and Optimization for Chemical Processes, Ministry of Education,Shanghai,200237 China)

【机构】 华东理工大学化工过程先进控制和优化技术教育部重点实验室；

【摘要】 针对标准粒子群算法的缺点,本文提出了双变异算子的混合粒子群优化算法(DDPSO)。该算法通过两次高斯变异控制算法进程,同时动态调节惯性权重。大概率的最差适应度变异对惰性粒子重新初始化,增加搜索空间。小概率的最优适应度变异加强最优解附近范围搜索,增加种群多样性,通过对Benchmark函数的测试结果表明:DDPSO算法确保了全局和局部搜索性能的动态平衡,在收敛性和稳定性上均明显优于标准粒子群算法(LWPSO)和杂交粒子群算法(HPSO)。更多还原

【Abstract】 Aiming at the disadvantages of the standard particle swarm optimization(PSO),a hybird particle swarm optimization algorithm based on dual mutation(DDPSO)was proposed.Twice Gaussian mutation controlled the algorithm process.Inertia weight was adjusted dynamically.Big probability of worst fitness value mutation enhanced global exploration abilities.Little probability of best fitness value mutation maintained population diversity and jumped out of the local optimal point.The results on Benchmark functions show that DDPSO keeps balance between global and local search,and has great advantage of convergence property and robustness compared with standard PSO(LWPSO)and hybrid PSO(HPSO)algorithms.更多还原