一种基于校正因子的自适应简化粒子群优化算法  

Adaptive simplified particle swarm optimization algorithm based on correction factor

作　　者：李鑫滨[1] 马阳[1] 鹿鹭 

机构地区：[1]燕山大学工业计算机控制工程河北省重点实验室,河北秦皇岛066004

出　　处：《燕山大学学报》2013年第5期453-459,共7页Journal of Yanshan University

基　　金：国家自然科学基金资助项目(61172095);河北省自然科学基金资助项目(F20122203138)

摘　　要：针对已有粒子群算法中全局搜索和局部搜索存在盲目性和滞后性以及粒子的早熟收敛等问题,提出了一种基于校正因子的自适应简化粒子群优化算法。该算法在简化粒子群算法基础上,以粒子间平均粒距大小作为触发条件,对惯性权重、平均个体极值和全局极值进行自适应扰动。校正因子可以根据当前粒子群个体信息和全局信息自适应调整,从而完成对当前粒子状态及时准确的更新,最终使粒子可以准确而快速的找到全局最优解。对3种典型测试函数的测试结果表明该算法具有较高的全局和局部搜索能力、能够有效地避免算法陷入局部极值,是一种实用且高效的粒子群改进算法。To overcome the problems of blindness and hysteresis during the global and local search, as well as the premature con- vergence shortcoming, which are in the pre-existing particle swarm optimizer algorithm, an adaptive simplified particle swarm op- timization algorithm based on the correction factor is put forward in this paper. The proposed algorithm based on the simplied particle swarm optimization algorithm regards average-distance-amongst-points as the trigger condition and does the adjustment to inertia weight, the average individual extremum and global extremum. The correction factor can adapt itself according to the personal and global information of presennt particle swarm, thus updates the present particle timely and accuratly so that it can help the particles find the golbal optimal solution quickly. The experiments results of three typical testing function present that this new algorithm owns high global and local search ability and is able to effectively avoid particles trapped into local optimal solution. In conclusion, it＇s a practical and effective improved partical swarm algorithm.

关 键 词：简化粒子群 平均粒距 校正因子 自适应 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]