基于改进粒子群算法的Pareto多目标输电网规划

Power transmission network planning of pareto multi-objective based on improved PSO

针对标准粒子群算法易收敛到局部最优的缺点,本文对粒子群算法做出了以下几点改进:首先,在编码策略上采用一种保证网络连通性的编码方式,有利于保持种群的分散性;其次,采用了一种改进的粒子速度更新公式,即在粒子群算法速度更新公式的基础上,加入一个平均极值项,使得各粒子能参考其它同伴的信息;另外,在算法迭代过程中加入变异操作,是使初始化失活粒子的位置和速度来保持种群多样性。在输电网扩展规划模型中引入了Pareto多目标模型,这种模型相对于单目标和加权多目标模型相比更具实际工程意义。算例结果表明,上述几个操作可以提高粒子群算法的收敛精度,使算法最终寻找到全局最优解,从而证明了改进粒子群算法的有效性.

For the shortage of standard particle swarm algorithm is likely to converge to local optimum,the following improvements are made on particle swarm optimization：firstly,apply a coding approach In the coding strategy to ensure network connectivity is conducive to maintaining the dispersion of population;Secondly,apply an improved particle velocity updating formula,that is to say add an average of extreme value items based on velocity update formula of PSO,so the particle can refer to other peer information;thirdly,join the mutation operation in the iterative process is to make the position of the initialize inactivation particle and velocity to keep the diversity.Introduce the Pareto multi-objective model in the transmission network expansion planning model,this model relative to the single-objective and multi objective model of weight compared to more practical engineering significance.Numerical results show that the number of operations can improve the accuracy of the convergence of PSO,the algorithm to find the global optimal solution,which proved that the improved particle swarm algorithm.