求解安全约束机组组合问题的邻域搜索外逼近方法  

An Outer Approximation Method Based on Neighborhood Search for Security Constrained Unit Commitment

作　　者：全然[1] 张晓菲 QUAN Ran;ZHANGXiaofei(College of Science,Henan University of Technology,Zhengzhou 450001,China)

机构地区：[1]河南工业大学理学院,郑州450001

出　　处：《河南科学》2018年第11期1699-1706,共8页Henan Science

基　　金：国家自然科学基金项目(71201049)

摘　　要：提出一种求解安全约束机组组合(security constrained unit commitment,SCUC)问题的邻域搜索外逼近(outer approximation based on neighborhood search,NS-OA)法. OA将SCUC问题分解为一系列混合整数线性规划(mixed integer linear programming,MILP)主问题和非线性规划(nonlinear programming,NLP)子问题,通过MILP主问题和NLP子问题的最优解来逼近SCUC问题的最优解.为克服迭代过程中MILP主问题规模大的不足,利用SCUC问题对应UC问题的最优解为中心来构造邻域,然后在此邻域内搜索MILP主问题的最优解.数值结果表明,所提邻域搜索能有效减小搜索空间,大大提高了算法的计算效率,所提NS-OA算法能有效求解大规模SCUC问题,具有良好的应用前景.This paper proposes an outer approximation method based on neighborhood search （NS-OA） for security constrained unit commitment （SCUC）. The OA decomposes the SCUC into a series of mixed integer linear programming （MILP） master-problems and nonlinear programming （NLP） sub-problems. Then, the optimal solution of SCUC is approximated by the solutions of the MILP master-problems and the NLP sub-problems. To overcome the weakness of the oversize scale of the MILP generated by OA, a neighborhood is built based on the solution of UC corresponding to SCUC to search the optimal solution of the MILP master problem. The numerical results show that the neighborhood search technique can reduce the search space efficiently and can improve the computing power greatly. Moreover, the presented NS-OA can solve large scale SCUC problems efficiently and have good application prospects.

关 键 词：安全约束机组组合问题 邻域搜索 外逼近法 混合整数非线性规划 混合整数线性规划 非线性规划 

分 类 号：TM71[电气工程—电力系统及自动化]