Multi-Strategy Improved Arithmetic Optimization Algorithm
Chuangsen Xie, Longri Wen, Huiqi Zhao, Insuk Sohn, Jooyong Shim, Changha Hwang, Jianye Zhuo, Xian Song · 2023
The Arithmetic Optimization Algorithm (AOA) represents a heuristic technique for optimization, exhibiting powerful local search capabilities. Nonetheless, it struggles with limited global search capacity and is prone to becoming stuck in local optima. Within this investigation, our team presents a Multi-strategy Enhanced Arithmetic Optimization Algorithm (MI-AOA) to surmount these constraints. The MI-AOA incorporates several improvement strategies: Enhancement of the acceleration function MOA to balance local and global search capabilities; improvement of the exploration phase using the Cauchy exploration strategy, which replaces the multiplication and division operators, providing a wider search range; and introduction of a random reverse learning strategy mutation to enhance population variety and avoid getting stuck in local optima. The proposed MI-AOA algorithm framework integrates these strategies, aiming to improve AOA's ability to escape local optima and strengthen global optimization capabilities. The framework consists of an initialization step, an iteration loop with exploration and exploitation stages, a best solution update process, and a mutation strategy. The results demonstrate that the MI-AOA is capable of achieving better global optimization performance compared to the standard AOA.