贪婪随机自适应搜索法在电网规划中的应用  

Application of Greedy Randomized Adaptive Search Procedure in Transmission Network Planning

作　　者：金华征[1] 程浩忠[1] 奚珣 夏夷 奚增辉 沈晓岚 

机构地区：[1]上海交通大学电气工程系,上海200240 [2]上海市东供电公司,上海200122

出　　处：《上海交通大学学报》2006年第4期563-567,共5页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金资助项目(50177017);上海市重点科技攻关计划资助项目(041612012)

摘　　要：基于贪婪随机自适应搜索法(GRASP)能有效地解决电网规划的组合优化问题,其每一次迭代包含构造和局域搜索两阶段.在构造阶段,以改进线路综合有效性指标为贪婪函数,采用比例法形成限制候选列表,并随机选择元素构成可行方案.根据线路的实际效用比剔除可行方案中的冗余线路,为局域搜索提供一个简练经济的初始网络,并采用1-1交换产生邻居集来拓展搜索空间寻找局优解.所有迭代中的最好方案作为电网规划问题的最优解.对于网络规模增大所引起的“维数灾”,提出5种邻居删减技术来提高算法的解算速度.算例分析证明了该方法的可行性、有效性.Greedy randomized adaptive search procedure （GRASP） is a high-efficient heuristic randomly iterative algorithm, which can effectively solve the combinatorial optimization problem of transmission network planning. Each iteration includes a construction and a local search phase. In the construction phase, the improved line integrated validity index is introduced as the greedy function, and a restricted candidate list （RCI.） is constructed by rule of proportion, from which elements are randomly chosen to form a feasible solution. The redundant lines in the feasible solution are eliminated according to the actual effect of the lines, which provides an efficient and economical initial network for local search, and the local optimal solution is sought by using 1--1 exchange method to bring neighbor set to extend search space. The best solution over all GRASP iterations is regarded as the global optimal one of the transmission network planning. As to the ＂dimension tragedy＂ brought by the network scale augmentation, five neighbor-pruning techniques are presented, which improves the computing speed of this algorithm. The simulation results show the availability and efficiency of this algorithm.

关 键 词：电网规划 贪婪随机自适应搜索方法 随机过程 贪婪函数 限制候选列表 

分 类 号：TM715[电气工程—电力系统及自动化]