Ding Feng, Dengao Li, Xiaojie An, Yu Zhou
Abstract Against the backdrop of the global energy transition toward decarbonization and digitalization, the high penetration of renewable distributed energy resources has made microgrid optimal scheduling face complex challenges such as multi-objective conflicts and source-load variability. To address issues in existing research, such as non-standardized algorithm benchmarking and models detached from engineering practice, this paper conducts a comparative study of multi-objective meta-heuristic algorithms spanning a diverse set of representative and state-of-the-art methods. A three-dimensional objective function including economic cost, carbon emissions, and tie-line power fluctuations is constructed, integrating a baseline-based quota and ladder-type carbon trading mechanism. Considering physical device constraints, a high-dimensional microgrid dispatch model is established. A test suite covering various meta-heuristic algorithms is designed, and a General Multi-Objective Benchmarking Framework (GMO-BF) is proposed to unify constraint handling, archive management, and performance evaluation standards. Furthermore, the robustness and sensitivity of algorithm performance are evaluated under multiple uncertainty levels through Monte Carlo simulations. Simulation results show that the proposed framework can effectively support fair comparison of algorithm performance, and the selected optimization algorithms can achieve a good trade-off among multiple objectives. While ensuring the economical and efficient operation of microgrids, it maintains acceptable carbon emission levels and grid stability, providing reliable technical support for microgrid energy management.