基于混合遗传算法的变电站选址定容

针对变电站优化规划这种大规模组合优化问题,提出了一种结合遗传算法与交替定位分配算法的混合遗传算法(GA-LA).该算法采用新型的三维编码策略,同时包含新建站的数量、站址和站容信息,并设计了适用于此新型编码的交叉算子和变异算子,以实现站址、站容的优化.其中,针对站址优化子问题,GA-LA算法将交替定位分配算法与遗传算法结合,在标准GA算子之后增加了一个LA算子,由GA算子进行种群中的全局广度搜索,LA算子进行染色体中的站址局部深度搜索,可实现无待选站址的自动寻优.算例结果表明,该方法具有较好的站址站容寻优能力和收敛性能,能满足实际电网中大规模变电站规划的需求.

Substation Locating and Sizing Based on Hybrid Genetic Algorithm

A novel hybrid genetic algorithm (GA-LA) based on alternative location-allocation and genetic algorithm is proposed to optimize the locations and sizes of distribution substations without candidate locations. The algorithm adopts a three-dimensional coding strategy including numbers, locations and sizes of new substations. It also designs the crossover and mutation operator adaptive to the strategy. For location sub-problem, the GA-LA algorithm combines alternative location-allocation and genetic algorithm together and after GA operator's global searching in the population, the LA operator for local optimization in the chromosomes is developed. The results demonstrate that GA-LA algorithm has better global searching performance and convergence property than traditional algorithm. The method proposed has a promising application in large-scale practical problem.