摘要
在对机组组合问题建立混合整数规划模型时,其最小开停时间约束的分离不等式的"松紧"程度直接影响求解算法的性能。首先从几何角度给出判断分离不等式优劣的直观判据(即松弛问题的可行域的大小),然后分析了常用最小开停时间约束的各种分离不等式,得到理论上最"紧"的1组机组最小开停时间分离不等式,最后通过仿真算例和实际算例验证了最"紧"分离不等式的正确性。
The "loose and tight" extent of valid inequalities of minimum start-up and trip-out time constraints in the mixed integer programming model for unit commitment problem directly impacts the performance of the algorithm to solve the model. Firstly, the authors give an intuitive criterion to judge the superior and inferior of valid inequalities in the viewpoint of geometry, i.e., the scale of the slack problem's feasible domain; then various valid inequalities of commonly used minimum start-up and trip-out time constraints are analyzed, thus the valid inequality of minimum start-up and trip-out time of one unit, which is theoretically most "tight'~, is obtained; finally, the correctness of the most "tight" valid inequality is verified by simulation results of IEEE RTS 96 system and calculation of actual example.
出处
《电网技术》
EI
CSCD
北大核心
2011年第5期82-89,共8页
Power System Technology
基金
国家电网公司科技项目(SG0874)
"十一五"国家科技支撑计划重大项目(2008BAA13B06)~~
关键词
混合整数规划
最小开停时间约束
多面体理论
多胞形
分离不等式
边界面
mixed integer programming
minimum on and off time constraints
polyhedral theory
polytope
valid inequalities
facet