云模型和混沌粒子群算法的多目标无功优化

为了克服基本粒子群算法易陷入局部最优值和后期收敛速度慢的不足，提出一种基于云模型的自适应粒子群算法。该算法首先采用混沌优化策略对粒子群进行初始化，增加粒子取值的多样性；其次根据粒子的适应度值将种群中的粒子分成靠近最优值、较靠近最优值和远离最优值3个子群，并分别采取不同的惯性权重生成策略进行处理，其中较靠近最优粒子子群的惯性权重由正态云发生器动态自适应调整，摆脱算法陷入局部最优值束缚；最后在迭代后期通过正态云算子实现粒子的变异操作，使算法后期快速收敛到最优解。对标准IEEE30节点系统和IEEE118节点系统进行测试仿真，结果表明了文中算法解决多目标无功优化的有效性。

Multi-objective reactive power optimization based on cloud model and chaos particle swarm optimization

In order to overcome the of the shortcomings particle swarm algorithm , such as slow convergence and easy stagnation in local optima , a new scheme based on cloud model and chaos particle swarm optimization algo-rithm is introduced to complete multi-objective reactive power optimization .Firstly, the chaos particle swarm initial-ization strategy is used to increase the diversity of the particle values .Secondly , according to the fitness value of the particles , the populations of particles are divided into three subgroups , ie the near optimal values , closer to the op-timal value and away from the optimal value .The algorithm can avoid trapping in local optimum value through using normal cloud generator algorithm to finish adaptive dynamic adjustment .Finally, in later stage of iteration the muta-tion operation of the particles is finished by the normal cloud particle operator and the algorithm can quickly con-verge to the optimal solution .Using standard IEEE 30 node system and IEEE 118 node system , the experiment re-sults show that the proposed scheme is effective in solving multi-objective optimization of the reactive power .