Savannah Bengal Tiger Optimization (SBTO): A Novel Metaheuristic Algorithm for Constrained Optimization Problems

This paper introduces a novel optimizer based on animal survival experiments called Savannah Bengal Tiger Optimization (SBTO). Inspired by the survival behavior of Bengal tigers on the African savannah, SBTO aims to address continuous complex constrained optimization problems. SBTO simulates the group hunting behavior of Bengal tigers and integrates the support of Kalman filters, employing three strategies: prey search, stealth approach, and hunting. The prey search strategy reflects SBTO's exploration capabilities, while the stealth approach and hunting strategies primarily demonstrate its exploitation capabilities. Compared to other metaheuristic algorithms, SBTO has an advantage in population distribution, maintaining good exploration performance while performing exploitation, which helps the algorithm escape local optima in a timely manner. Finally, SBTO was experimentally evaluated against 10 popular algorithms and recently proposed algorithms on CEC2017, CEC2020, CEC2022 test functions, and 9 engineering problems. The results indicate that SBTO achieved the best fitness ratio of 27/30, 8/10, and 8/12 in the test functions, with Wilcoxon rank-sum tests showing significance proportions of 260/300, 89/100, and 104/120, respectively. In the 9 engineering problems, SBTO obtained the best average and optimal fitness in 7 problems, demonstrating exceptional performance in constrained optimization problems and complex multi-modal functions. The source code for SBTO is publicly available at https://www.mathworks.com/matlabcentral/fileexchange/172500-sbto.