Abstract
Enterprise Development Optimizer (EDO) is a meta-heuristic algorithm inspired by the enterprise development process. Although EDO is effective in the optimization field, it suffers from issues such as premature convergence and unequal exploration-exploitation ratio. These shortcomings restrict the performance of the algorithm in some complex problem. This research proposes an improved EDO, called LMEDO, in which EDO is integrated with incorporates time-phase based switching strategy, economy-driven guided based learning strategy and spatial selectivity-based selection strategy to improve convergence rate, stability, and search effectiveness. Among these strategies, the time-phase based switching strategy allows EDO to better apply different search strategies and enhances the search capability. Economy-driven guided learning-based strategy helps EDOs absorb valid information from dominant groups, which in turn improves the quality of the entire population. The spatial selectivity-based selection strategy achieves a balance between exploitation and exploration capabilities. To validate the performance of LMEDO, an extensive evaluation of the CEC 2018 test suite and five engineering optimization problems was performed. Parameter sensitivity analysis assisted LMEDO in determining the optimal parameter settings. Ablation experiments confirmed the effectiveness and compatibility of the improved strategies. The superiority of LMEDO is validated by comparing it with state-of-the-art algorithms such as LSHADE-SPACMA, APSM-jSO, and GLS-MPA. LMEDO received an average ranking of 2.5862 on the CEC2018 test suite and obtained a result of 1161/94/143 (+/=/-) on the Wilcoxon rank sum test. In addition, engineering design optimization problems are investigated to further demonstrate the reliability and flexibility of LMEDO. In conclusion, LMEDO is a promising variant of metaheuristic algorithms and is effective and accurate for solving complex problems.