Abstract: Reactive load history information cannot be effectively applied by traditional reactive compensation classification method, so it exists the over-compensation and lack-compensation phenomenon easily. It proposes a optimization model for effective use of reactive load history information to determine the classification capacity. The model is solved using the improved ant colony algorithm. The pheromone is corrected in time by setting a pheromone threshold. Searching in vertical and horizontal way, the efficiency of ants’ search is improved. Solution in that algorithm is effectively protected from local optimum, and the efficiency is increased several times.

Key words: ant colony algorithm, reactive compensation, compensation classification, pheromone, correction factor, optimal solution

摘要： 传统确定无功分级补偿容量的方法不能有效利用负荷历史信息，容易出现过补或欠补现象。建立了有效利用历史无功负荷来求解无功补偿分级容量的优化模型，采用蚁群算法求解，对蚁群算法进行了改进。通过设定信息素的修正阈值，适时对信息素进行修正；通过纵向和横行的搜索方式，提高蚂蚁搜索的效率；算法能更好地避免陷入局部最优，执行效率数倍提高。

关键词: 蚁群算法, 无功补偿, 分级补偿, 信息素, 修正因子, 最优解