| 引用本文: | 李佩杰,万海涛,赵晓慧,等.定制化求解机组组合混合整数线性规划模型的固定—推断法[J].电力系统保护与控制,2023,51(2):12-22.[点击复制] |
| LI Peijie,WAN Haitao,ZHAO Xiaohui,et al.A customized fix and implicate method for mixed integer linear programming models ofunit commitment problems[J].Power System Protection and Control,2023,51(2):12-22[点击复制] |
|
| 摘要: |
| 为了突破机组组合算法的自主可控问题，基于开源混合整数线性规划求解器CBC，提出一种快速获取机组组合问题可行解的固定—推断法。首先将机组组合模型转换为推断标准模型，然后按重要性对所有整数变量进行排序。并利用约束违反函数依次确定整数变量的值，实现整数变量的固定，利用约束关系推断出与其相关的整数变量值。最后经过多轮的固定—推断可以实现所有整数变量的取值，从而求解一个线性规划问题即可得到各机组的出力。仿真结果表明，所述算法能有效求解大规模机组组合问题，可在更短时间内获取质量较好的可行解。与CBC求解器结合，能显著提升CBC求解器对于机组组合问题的求解效率。此外，所述算法还具备在其他求解器上进行定制的潜力。 |
| 关键词: 机组组合 混合整数线性规划 CBC求解器 固定—推断法 |
| DOI：10.19783/j.cnki.pspc.220548 |
| 投稿时间：2022-04-18修订日期：2022-08-25 |
| 基金项目:国家自然科学基金项目资助(51967002，52267006) |
|
| A customized fix and implicate method for mixed integer linear programming models ofunit commitment problems |
| LI Peijie,WAN Haitao,ZHAO Xiaohui,WEI Hua,YANG Ming |
| (1. Guangxi Key Laboratory of Power System Optimization and Energy Technology (Guangxi University), Nanning 530004,
China; 2. College of Electronic Information, Guangxi University for Nationalities, Nanning 530006, China; 3. Key
Laboratory of Power System Intelligent Dispatch and Control (Shandong University), Jinan 250061, China) |
| Abstract: |
| To obtain an independent and controllable method for unit commitment (UC) , based on the open source mixed integer linear programming (MILP) solver CBC, this paper proposes a fix and implicate (F&I) method to fast obtain the feasible solution of UC problem. First, the implicate standard model is obtained by transforming UC model, and all integer variables are sorted by their importance. Then, the integer variables are sequentially fixed to values determined by the constraint violation function. The values of other related integer variables are implicated by constraint relation in each round of fixing. Through some rounds of fixing and implicating, all integer variables are fixed quickly. Finally, the power output of each unit can be obtained by solving a linear programming problem. The simulation shows that the proposed method can effectively solve large-scale UC problems and obtain better feasible solution in a short time. Combined with the proposed method, CBC solver can solve UC problems more effectively. In addition, the F&I method also has the potential for customization on other solvers.
This work is supported by the National Natural Science Foundation of China (No. 51967002 and No. 52267006). |
| Key words: unit commitment mixed integer linear programming CBC solver fix and implicate method |
