Beta Random Restart Strategy-Based Remora Optimization Algorithm for Global Optimization

The Remora Optimization Algorithm (ROA) is a meta-heuristic algorithm that imitates the foraging behaviors of Remora. Its main idea lies in simulating the mechanism of switching hosts during the foraging process of remora. Due to the randomness of remora host selection, ROA frequently gets trapped in local optima, which slows down its convergence speed. To develop a more robust algorithm, this paper simulates the exploration and elimination mechanism in the biological evolution process and improves the random restart strategy with “prior” properties. Beta Random Restart Strategy-based Remora Optimization Algorithm (BROA) is proposed to realize global optimization. The proposed Beta-random restart strategy ranks individuals by fitness and restarts the less fit half. The Cauchy strategy is used to expand the search area and enhance the ability to escape from local optima. This paper meticulously assesses the performance of BROA using five comparison algorithms. Firstly, the optimization capabilities of BROA are assessed through CEC2020 tests. The Wilcoxon test assesses the difference between BROA and five different algorithms. Finally, Cantilever Beam Design problem is used for testing the practicability of BROA. Comprehensive results show that BROA performs best in CEC2020 and the cantilever beam design problem compared with five different optimization algorithms.