基于地理信息因子及自适应小生境微分进化算法的变电站规划

摘要

A GIFANDE(Geography Information Factor and Adaptive Niche Differential Evolution) algorithm is proposed for substation sizing and locating of modern urban power grid. A comprehensive planning model is established based on the interval analytical hierarchy process method,which considers the expenditure of substation construction and operation,as well as different geographic information factors,such as land nature, transportation condition, flood control & drainage, geographical features, construction conditions, etc. The adaptive niche differential evolution algorithm is adopted in optimization process, which adjusts the individual adaptive value,speeds up the convergence by elimination,and adaptively adjusts the niche radius according to the intervals between individuals to improve the capability and efficiency of global optimal search. Practical example shows the effectiveness of the proposed algorithm in the substation planning of urban distribution network.%针对现代城市配电网变电站选址定容问题,提出一种充分考虑地理信息因子影响的自适应小生境微分进化算法.建立基于区间层次分析法,考虑用地性质、交通情况、防洪排水、地质地貌、施工条件等因素的地理信息因子和变电站建设、运行等费用的综合规划模型.在问题寻优过程中,在微分进化算法的基础上引入小生境中共享机制构成小生境微分进化算法,该算法改变个体适应度值,通过淘汰运算加快收敛速度,并根据个体间相对距离判断种群的聚集情况以自适应调整小生境半径,从而较大提高了算法的全局寻优能力和搜索效率.实际算例表明,所提算法能较好地解决城市配电网变电站规划问题.