An Efficient Optimization Method for Randomly Aggregating Unit Variables in SCUC

To expedite the achievement of crucial objectives such as carbon peaking and carbon neutrality, and to address the issue of limited efficiency in unit commitment in new energy systems reliant on traditional thermal power, we propose a highly effective optimization method for aggregating random unit variables. The proposed method acquires binary variable information from the original problem by creating several smaller-scale problems, which is achieved by randomly amalgamating the decision variables from the original problem and solving a smaller-scale model that closely emulates the original problem's objective function and constraints. Through the utilization of 0–1 variables, we can effectively reduce variable dimensions and generate heuristic solutions, thereby enhancing the efficiency of addressing security-constrained unit commitment problems. Numerical results demonstrate that the proposed algorithm outperforms the direct solution method in terms of rapidly discovering high-quality feasible solutions. Furthermore, the algorithm also exhibits a swifter rate of improvement in the lower bound. This enhancement significantly boosts the efficiency of solving optimization problems and effectively addresses energy system scheduling.