Enhanced stellar oscillation algorithm for optimal design of inudstrial components

Abstract

This paper introduces the Modified Stellar Oscillation Optimizer (MSOA) for constrained industrial design problems. The base optimizer, which models the pulsation and oscillatory behavior of stars, is hybridized with chaotic maps. MSOA is tested on five constrained engineering problems: a speed reducer, a cantilever beam, a step-cone pulley, a rolling element bearing, and a vehicle structural component. On the cantilever beam, MSOA reaches the global feasible optimum f  = 1.33996, tying with CSA, FOX, GWO, AVOA, and SMA. On the speed reducer and step-cone pulley, MSOA reaches the global feasible optima of 2,996.348 and 16.0899, matching SFOA and BBO and improving on the other tested algorithms by 0.1–4 %. For the rolling element bearing and the automobile component, MSOA produces better objective function than several other recent metaheuristics. The chaotic hybridization helps MSOA remain in the feasible region on problems where most rivals converge to penalty-corrected local minima.