Multi-strategy enhanced Human Evolution Optimization Algorithm and its engineering applications

To address the limitations of the original Human Evolution Optimization Algorithm (HEOA) in global search capability, adaptability, and convergence accuracy, this paper proposes a multi-strategy Enhanced Human Evolution Optimization Algorithm (EHEOA). The algorithm integrates a polynomial Lévy flight strategy to enhance global exploration, employs a linearly decreasing step-size mechanism for refined local exploitation, introduces dynamic role allocation to improve search flexibility, and designs novel operators such as diamond vector crossover and Gaussian perturbation to maintain population diversity. To validate the algorithm’s performance, systematic experiments were conducted on a total of 42 benchmark functions from the CEC2017 and CEC2022 test sets. On the 30 complex functions of CEC2017, EHEOA achieved theoretical optimal solutions on 10 functions, with an average ranking of 2.27, significantly outperforming mainstream algorithms such as WOA and HHO. On the more challenging CEC2022 test set, EHEOA obtained optimal results on 10 out of 12 functions. Overall, EHEOA achieved the best solutions on 47.6% of the test problems, demonstrating excellent convergence speed and stability across unimodal, multimodal, and composite function optimization. The Wilcoxon rank-sum test further confirmed that EHEOA exhibits statistically significant advantages in most optimization tasks. Experimental results show that EHEOA effectively balances exploration and exploitation through its multi-strategy cooperative mechanism, demonstrating strong adaptability and competitiveness in various complex optimization scenarios, thereby providing a new solution for engineering optimization problems.