UNIT COMMITMENT WITH AN EFFICIENT FORMULATION AND ALGORITHM BASED ON IDENTIFICATION METHOD FOR VALID OPTIMIZING SPACE

摘要

Unit Commitment is one of the most important issues in power system operation, and solving this problem with mature optimizing algorithms has been the general trend now. However, too many integers which are needed to be branched are the biggest trouble to obtain the optimal solution in limited time. Different from most of the published papers, which mainly focused on the improvement of a certain optimizing algorithm, a novel notion of optimizing space of unit commitment and its scientific identification method are proposed in this paper from the perspective of model optimization. Based on the analysis on both units and binary variables of each time interval, this method can identify the invalid integer variables which need not to be branched, and turn them into continuous variables. Besides, in response to the failure of the traditional formulation of mixed integer linear programming, a novel one based on the valid optimizing space is proposed. Based on the work in this paper, the optimizing range is narrowed efficiently, and the calculating efficiency is improved greatly, under the premise of guaranteeing feasibility and optimality of original problem. The effectiveness of this method is demonstrated by the theoretical analysis and numerical experiments.