并行自适应粒子群算法在电力系统无功优化中的应用

针对传统粒子群优化算法"早熟"与后期收敛速度慢的缺点,提出了一种基于并行自适应粒子群优化算法的电力系统无功优化方法。该方法首先将初始种群随机划分成N个子群,然后分别在各子群中以所提方法寻优,从而实现了算法的并行计算。为避免各子群陷入局部最优解,采用二值交叉算子使各子群间的信息共享并更新相关粒子位置,保证了算法的全局搜索能力并维持了种群的多样性。同时,各子群寻优过程中,根据利己、利他及自主3个方向对当前搜索方向自适应更新,提高了算法的收敛速度。将所提出算法在IEEE 30节点系统上进行了仿真验证,结果证明了并行自适应粒子群算法用于无功优化的可行性和有效性。

Application of Parallel Adaptive Particle Swarm Optimization Algorithm in Reactive Power Optimization of Power System

There are defects in traditional particle swarm optimization(PSO) algorithm,i.e.,its prematurity and slow convergence speed in the late evolutionary phase.For this reason,a method for power system reactive power optimization based on a parallel adaptive PSO(PAPSO) algorithm is proposed.Firstly,the initial population is divided into N subpopulations stochastically;then the search in each subpopulation is performed individually by the proposed method,thus the parallel calculation of the adaptive PSO algorithm is implemented.To avoid the search in subpopulations falls into local optimal solution,the two-value crossover operator is led in to exchange the information among subpopulations and update the position of related particles,thus the global search ability of the algorithm is ensured and the diversity of population can be kept.During the search process in each subpopulation,current search direction is adaptively updated in accordance with egotistic direction,altruistic direction and pro-activeness direction to improve the convergence speed of the algorithm.Simulation results of IEEE 30-bus system show that as for reactive power optimization the proposed algorithm is feasible and effective.