摘要
基本粒子群优化算法每个粒子代表一个可行解,通过粒子间的协作来获得最优解。考虑粒子间协同作用,引入Gaussian核函数研究基于区域影响的粒子群算法(GPSO)。为了充分利用粒子群算法的快速全局收敛性和模拟退火算法能够跳出局部最优陷阱的优点,得到高精度的最优解,将GPSO算法与模拟退火算法相结合,研究了一种新的混合粒子群算法。混合算法在GPSO算法处于停滞状态时,于搜索到最优位置用模拟退火算法继续寻找最优解。数值实验结果表明,新混合算法兼顾了GPSO和模拟退火算法的优点,具有收敛速度快、搜索精度高、鲁棒性好等特点。这说明文中的混合算法不失为一种有效的进化算法。
Basic Particle Swarm Optimization （PSO） algorithm, of which each particle represents a feasible solution, obtain the best solution through the cooperation between particles. Considered the effect between particles, the PSO based on the area of influence is studied by introducing a Gaussian kernel function. Combining GPSO with Simulated Annealing （SA） algorithm, study a new hybrid optimization algorithm to get high precision of the optimal solution. The hybrid algorithm applies SA in the best position found by the GPSO at the stagnation of evolution progress, and continues to search for the optimal solution. The experimental results show that the new hybrid optimization algorithm takes into account advantages of both GPSO and SA, has advantages at convergence speed, convergence accuracy and robustness. This shows that the hybrid algorithm can be regarded as a kind of effective evolutionary algorithm.
出处
《计算机技术与发展》
2013年第7期26-30,共5页
Computer Technology and Development
基金
重庆市自然科学基金资助项目(CSPC
2005BB2197)
重庆大学"211工程"三期创新人才培养计划建设基金资助项目(S-09110)
