Enhanced artificial ecosystem-based optimization for global optimization and constrained engineering problems

Abstract Artificial ecosystem-based optimization (AEO) is a nature-inspired intelligent optimization algorithm that has been widely applied to various real-world optimization problems. However, AEO has several limitations, including slow convergence and difficulty in escaping from local optima. To address these drawbacks, this study proposes an enhanced variant of AEO called enhanced artificial ecosystem-based optimization (EAEO). First, Latin hypercube sampling is introduced to achieve uniform population initialization. Then, a quadratic interpolation mechanism is embedded to accelerate convergence and improve accuracy. Finally, an adaptive neighborhood search inspired by animal migration behavior is designed to help to jump out of local optima. The performance of EAEO is evaluated using twenty-three benchmark functions and the CEC2017 test suite. Experimental results indicate that EAEO outperforms the original AEO and other comparison algorithms in terms of accuracy and stability. The proposed EAEO is applied to solve four engineering optimization problems. The results demonstrate the superiority of EAEO in addressing practical problems.