基于二进制微分进化算法和目标函数分解的大规模机组组合求解

Large-scale unit commitment solution based on binary differential evolution algorithm and objective function decomposition

为了避免在机组组合求解过程中将机组启停计划和负荷经济调度两者形成内外双层嵌套求解,从而导致计算比较耗时的问题,引入启运机组的总平均燃料成本和系统旋转备用剩余量这2个可调节的子目标,将传统的机组组合模型分解成2个独立的优化目标,构建了一种基于目标函数分解的二阶段可独立求解的机组组合模型。采用一种改进的二进制微分进化算法对第一阶段的机组启停计划目标进行求解,对每个代表机组启停状态的个体编码采用机组最小启停时间约束、旋转备用约束、机组去组合等处理机制,有效保证了每个解的有效性并缩小了算法的搜索空间。根据求解得到的机组启停状态,采用半定规划法求解第二阶段的负荷经济调度目标。采用经典的测试算例验证了所提方法在大规模机组组合求解中的有效性。

In order to avoid the problem of time-consuming caused by the internal and external two-level nested solution of unit start-stop schedule and load economic dispatch in the solving process of UC(Unit Commitment),two adjustable sub-objectives,namely the total average fuel cost of start unit and the system’s spare spinning reserve capacity,are introduced to decompose the traditional UC model into two independent optimization objectives,then a two-stage UC model based on the objective function decomposition is proposed,in which the sub-stage model can be solved independently. An improved BDE(Binary Differential Evolution) algorithm is used to solve the unit start-stop schedule objective in the first stage. The handling mechanisms such as the minimum unit start-stop time constraint,the spinning reserve capacity constraint, the unit de-commitment,and so on,are applied to each individual coding that represents the start-stop status of unit,which can effectively guarantee the validity of each solution and reduce the search space of the algorithm. According to the solved start-stop status of unit,the semi-definite programming method is used to solve the load economic dispatch objective in the second stage. The effectiveness of the proposed method in solving the large-scale UC is verified by a classical test example.