The Graduate Student Evolution Algorithm Based on the Improvement of Mental Health and Ideological and Political Education for Global Optimization and Practical Problems

The Graduate Student Evolutionary Algorithm (GSEA) provides a simple population-based optimization framework, but its performance may be limited by unreliable mentor selection, insufficient adaptation of learning behavior, and inadequate utilization of historical search information. From the perspective of symmetry-related search coordination, maintaining an appropriate balance between exploration and exploitation while accommodating the asymmetric contributions of different individuals and search experiences is essential for effective optimization. To address these challenges, this study proposes an Improved Graduate Student Evolutionary Algorithm, termed IGSEA, for global optimization and grid-connected microgrid scheduling. Three complementary mechanisms are incorporated into the original framework. First, a mental-resilience-based adaptive learning regulation strategy adjusts individual search behavior according to relative fitness and stagnation information, facilitating a dynamic balance between global exploration and local exploitation. Second, a collective value-guided mentor–peer collaboration strategy integrates the global best solution, rank-weighted elite information, population consensus, and peer differences to construct more reliable search guidance while coordinating individual and collective information. Third, a practice–reflection-based knowledge feedback strategy employs asymmetric feedback from successful and unsuccessful search movements, retaining beneficial directions and using unsuccessful updates to generate corrective information. Through these mechanisms, IGSEA seeks to coordinate competing search requirements without imposing artificial symmetry on the optimization landscape. The proposed algorithm is evaluated on the CEC2017 benchmark suite under 10-, 30-, 50-, and 100-dimensional settings. Numerical results, convergence curves, boxplots, Friedman rankings, and Wilcoxon signed-rank tests demonstrate its competitive solution accuracy and robustness. IGSEA obtains Friedman mean ranks of 1.47, 1.33, 1.43, and 1.87 under the four dimensional settings, respectively. Furthermore, IGSEA is applied to day-ahead scheduling of a grid-connected microgrid comprising renewable energy sources, dispatchable generators, battery energy storage, and utility-grid interaction. It achieves a best scheduling cost of 1275.65 and an average cost of 1393.56, reducing the average cost by 9.42% compared with CBSO and 19.15% compared with the original GSEA. These results indicate that IGSEA provides a competitive optimization approach by coordinating adaptive search regulation, collective guidance, and asymmetric knowledge feedback in numerical and constrained engineering optimization problems.