Performance of multi-strategy optimized mayfly optimization algorithm for location selection of university reimbursement form submission terminals.

The Mayfly Optimization Algorithm (MOA) is a swarm intelligence algorithm with competitive search capability, but it may suffer from premature convergence and unstable late-stage exploitation. This study proposes a multi-strategy optimized mayfly optimization algorithm (MSMOA) to improve the overall optimization performance of MOA. MSMOA integrates Logistic chaotic initialization, a nonlinear adaptive weighting factor, Lévy-flight perturbation, and a PSO-guided learning mechanism to increase initialization randomness, adjust the search process, introduce long-range stochastic perturbations, and provide additional information-sharing guidance. The algorithm was evaluated on scalable benchmark functions under 30D, 50D, and 100D settings and two fixed-dimensional functions, and was compared with MOA, PSO, GWO, SSA, and DESMA over 30 independent runs. MSMOA achieved the best overall average rank among the compared algorithms and maintained competitive performance across high-dimensional settings. Two-sided Wilcoxon rank-sum tests supported the statistical reliability of the observed performance differences, and ablation experiments suggested that the four components jointly contributed to the overall performance improvement. Finally, a reimbursement-terminal location-selection case provided a preliminary illustration of the applicability of MSMOA to a simplified static facility-location problem.