基于自适应粒子群优化算法的机组组合  

A solution to particle swarm optimization algorithm with adaptive inertia weight for unit commitment

作　　者：常文平[1] 于海[2] 华大鹏[3] 

机构地区：[1]河南机电高等专科学校电气工程系,河南新乡453002 [2]河南电力公司新乡供电公司,河南新乡453002 [3]河南电力公司南阳供电公司,河南南阳453002

出　　处：《电力系统保护与控制》2009年第15期15-18,共4页Power System Protection and Control

摘　　要：机组组合是一个大规模、非线性混合整数优化问题,求解比较困难,为了提高粒子群算法的全局和局部搜索能力,提出一种惯性权值自适应调整的粒子群算法。该算法按照适应度的大小将粒子群分成两个子群,然后根据适应度的进化速度和进化停滞系数动态调整惯性权值。通过对典型函数的测试以及10台机组24小时的优化调度,计算结果表明该方法收敛精度较高。Unit commitment is a large-scale and mixed-integer non-linear programming problem. An adaptive inertia weight of particle swarm optimization algorithm （AWPSO） is presented to increase the global and local search. The population is divided into two sub-populations according to the value of fitness, and the inertia weight is formulated as a function of evolution speed and stagnate state. The algorithm is tested with well-known benchmark functions and the simulation results with systems of up to 10 units and 24-h scheduling horizon are presented. The experiments show that the convergence accuracy is increased.

关 键 词：粒子群算法 惯性权值 自适应 机组组合 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]