基于分层分枝定界算法的机组组合  

Unit commitment based on layered branch and bound algorithm

作　　者：谢国辉[1] 张粒子[1] 舒隽[1] 苏济归[1] 

机构地区：[1]华北电力大学电力市场研究所,北京102206

出　　处：《电力自动化设备》2009年第12期29-32,共4页Electric Power Automation Equipment

摘　　要：针对常规机组组合算法计算量大、计算精度不高的问题,提出了一种分层分枝定界算法。该算法采用分层求解策略,对传统机组组合模型进行线性化处理,进而基于线性规划算法求解松弛整数变量的线性化机组组合模型,通过取整策略形成初始分枝,作为分枝定界算法的上层;采用经典广度优先搜索算法的节点搜索策略,以及考虑负荷备用和机组启停时间约束并结合最大、最小边界的分枝策略,不断进行分枝、定界和剪枝获得下层问题的最优解。不同测试算例分析表明,所提出的分层分枝定界算法快速、稳定,能够考虑机组爬坡约束,可以在保证合理计算时间内有效提高求解精度。Aimed at the heavy computation and low precision of traditional uc（Unit Commitment） algorithms, the layered branch and bound algorithm is presented, which adopts the layered strategy to linearize the traditional UC model. The linear UC model of relaxed integer variable is then solved based on linear programming to form the initial branch consideration of the load reserve and on max - min border,the branching,bounding node search strategy to obtain the lower constraints effectively and security constraints, the as the upper layer of branch and bound algorithm. With the -off time constraints and combined with the branch strategy of and pruning are carried out continuously by the breadth-first layer optimal solution. Examples show that, considering the ramp layered branch and bound algorithm is fast and robust,and improves the solution accuracy in reasonable calculation time

关 键 词：机组组合 分层分枝定界算法 初始分枝 广度优先搜索 最大 最小边界 

分 类 号：TM73[电气工程—电力系统及自动化]