针对传统粒子群算法中收敛速度快但易于陷入局部最优等特点，将差异进化算法与粒子群算法相结合，提出了一种粒子群-差异进化混合算法。该算法在粒子寻优过程中除跟踪个体极值和全局极值外，还跟踪粒子差异进化产生的第三个值；同时，当粒子在某一维上的速度小于给定值时，将重新初始化该维度粒子速度。建立了无功优化数学模型，并将合算法应用到无功优化中。通过MATLAB编程对IEEE-30节点系统进行优化计算，并与遗传算法和粒子群算法比较，结果表明本文提出的算法应用于无功优化拥有较快的收敛速度和全局寻优能力，具有广阔的发展前景。%This paper presents a particle swarm optimization-difference evolutionary algorithm that aims to solve the flaws of easy plunging into local optimum and it is applied for reactive power optimization .In the algorithm each particle keeps track of the third value which is created by the mutation operator of DE algorithm besides the best previous position found so far by itself and the best previous position among all particles .Besides, the velocity is reinitialized and the dimension of the personal best position is mutated by mutation operator of DE algorithm if the dimension of one particle ’ s veloeity of PSO algorithm is smaller than the specified value .Through the establishment of reactive power optimization mathematical model , the proposed algorithm optimizes IEEE 30-bus system through the Matlab programming , and compares with particle swarm optimization and genetic algorithm .The optimized re-sults show that the proposed algorithm has better search capability and higher degree of convergence for reactive power optimization , and can control system according to the optimized results and hit the mark of decrease transmis-sion loss and improve the quality of voltage level .
展开▼
