蚁群算法在配电网重构的应用  

Application of Ant Colony System in Distribution Reconfiguration

作　　者：黄健[1] 张尧[1] 李绮雯[2] 

机构地区：[1]华南理工大学电力学院 [2]广东电网公司东莞供电局,东莞511700

出　　处：《电力系统及其自动化学报》2007年第4期59-64,共6页Proceedings of the CSU-EPSA

摘　　要：配电网络重构是一个非常复杂的大规模组合优化问题。蚁群算法作为一种现代启发式寻优技术,适合于求解组合优化问题,其主要特点是正反馈、分布式计算、易与其它算法结合以及富于建设性贪婪启发式搜索。对配电网络从图论拓扑结构上进行分析,将配网重构问题转化为求图的生成树问题,并以破圈法为基础得到快速而有效地求解图的生成树的方法。在应用蚁群算法求解配网重构问题时,通过首支路选择随机化和取消蚁群算法常用的启发值的方法,扩大算法搜索范围,使算法可以跳出局部最优化陷阱,改善算法的搜索效果。对IEEE 69网络的算例表明,该方法能以较少的计算量和较大的概率收敛于全局最优解。Distribution network reconfiguration for loss minimization is a complex, large-scale combinatorial optimization problem. As a new heuristic searching technique, ant colony algorithm is suitable for solving combinatorial optimization problem. ACO （ant colony optimization） has positive feedback and the ability of distributed computation, and it is easy to combine with the other algorithms and can do greedy heuristic search. After analyzing the distribution network in topology, the distribution network reconfiguration problem is converted into a problem of constructing spanning tree of the graph, which is solved by breaking-cycle-basis method. An improved ant colony optimization algorithm is proposed to solve the reconfiguration problem, which can expand the search extent and avoid search stagnation by selecting the first-branch randomly and canceling the heuristic value of network. Case study on IEEE 69-bus system proves that the proposed algorithm can obtain the global best solution with less computation time and higher probability compared to the conventional methods.

关 键 词：网络重构 蚁群算法 图论 破圈法 

分 类 号：TM72[电气工程—电力系统及自动化]