Modified Secretary Bird Optimization Algorithm with Multi-Strategy for Solving Optimization Problems
Ling-Long Tan, Yu-Ting Yu, Qi-Jian Wang · 電腦學刊 · 2025
The Secretary Bird Optimization Algorithm (SBOA) is an innovative meta-heuristic optimization technique, recently developed and inspired by the natural survival behaviors of secretary birds. In practical optimization scenarios, the SBOA, similar to other meta-heuristic algorithms, often suffers from slow convergence rates, susceptibility to local optima, and inconsistent exploitation capabilities. To improve the performance of the SBOA, a modified Secretary Bird Optimization Algorithm (MSBOA) with Multi-strategy Integration is proposed, which combines the SBOA algorithm, golden section search strategy, and differential evolution technique. The improvement includes two parts: During the exploration phase, the MSBOA employs a golden sine strategy to deepen and broaden local searches. To increase the convergence rate, differential evolution is applied to each population position update during the exploitation phase. To evaluate the effectiveness of the MSBOA in real-world scenarios, the approach under consideration was applied to the IEEE CEC-2021 benchmark functions and six real-world optimization problems. In addition, MSBOA was evaluated as a feature selection method using six standard datasets from the University of California, Irvine (UCI). The simulation results showed that MSBOA outperformed the SBOA and other metaheuristic methods in most benchmark functions, real engineering problems, and feature selection tasks. Thus, the results indicate that the proposed algorithm achieves highly competitive performance and outperforms other metaheuristic methods in solving optimization problems.