Advanced Ensemble Optimization Framework for Risk-based Energy Scheduling in Optimal Active-Reactive Power Dispatch

Abstract

Active and reactive power management and scheduling in power systems is critical for ensuring stable energy availability for residential, commercial, and industrial end-users. To address the scheduling of active/reactive power proportions, this paper proposes a novel Advanced Ensemble Optimization Framework named GIRCEDUMDA_AVOA_GWGwDVO. This hybrid methodology integrates the Artificial Vultures Optimization Algorithm (AVOA) for exploration, the Group-Weighted Genetic Algorithm with Dynamic Variable Optimization (GWGwDVO) for exploitation, and the Grouped Importance RCEDUMDA for statistical resource allocation. The strategy optimizes the parameter values through a phase-specific evaluation budget of a determined number of iterations (in a benchmarking that maximum accept 5000 iterations), significantly reducing performance indicators (risk based and expected cost). Simulation results on IEEE 30, 118, and Indian 62 bus systems demonstrate a convergence improvement of 37% and a computational efficiency increase of 42% compared to traditional methods. Extensive error analysis and numerical stability assessments confirm the robustness of the proposed approach.