Improved Dung Beetle Algorithm for Multi-Objective Environmental Economic Dispatch of Microgrid

With the widespread integration of renewable energy, microgrid environmental economic dispatch (EED) faces challenges such as uncertainties in wind and solar power outputs and multi-objective conflicts. This paper proposes a stochastic expected dispatch framework based on an improved multi-objective dung beetle optimization algorithm (MO-CLDBO). First, considering both wind–solar uncertainties and demand response, a Gaussian Copula function is employed to characterize the 24-h temporal correlations among wind speed, solar irradiance, and load, and typical scenarios are generated via Monte Carlo sampling and simultaneous backward reduction; a time-of-use demand response model is also introduced. Second, taking expected operational cost and environmental emission as dual objectives, three improvements are proposed to address the issues of uneven initial population, easy local convergence, and Pareto front collapse in the standard dung beetle algorithm: a Folded Two-Dimensional Modified Coupled Logistic-Sine Map (Folded 2D-MCLSM) is used to initialize a high-quality population, a non-dominated sorting mechanism is introduced, and a dynamic lens imaging backward learning strategy is designed. Finally, the proposed algorithm is compared with several classical algorithms in the mathematical model of microgrid optimal dispatch through 50 independent runs. Experimental results show that the improved dung beetle optimization algorithm achieves not only the lowest average operating cost, but also the best hypervolume (HV) indicator, demonstrating excellent comprehensive performance in multi-objective search convergence and solution set diversity.