电力网无功优化的一种新算法

A Kind of New Algorithm of Reactive Power Optimization of Electric Network

为了避免常规遗传算法容易出现的“早熟”现象,并提高算法的精度,在无功优化的进化过程中以种群散度为判断尺度将原种群分化为若干子种群,引入了伪并行遗传算法思想。对各子种群独立进行选择、交叉、变异等遗传操作,并在每代遗传操作结束后相互传递有利信息,使各子种群向着更优方向进化。通过对IEEE14节点系统进行测试,论述了所采用的模型及算法的合理性和可行性。

In order to avoid the easy occurence of “premature” phenomena in the simple genetic algorithm and to improve the algorithm precision,the original group can be differentiated into several subgroups with the group dispersion degrees as the judgement scales in the evolution process of reactive power optimization.The concept of pseudo-parallel genetic algorithm is introduced in carrying out the genetic operations such as selection, crossing and mutation of each sub-group independently.After the completion of genetic operation of each generation,the favorable information can inter-disseminate with each other,whereby making each sub-group evolve toward the optimal direction.Through testing IEEE 14-node system,this paper deals with the rationality and feasibility of the model and algorithm adopted.