Chaotic coyote algorithm applied to truss optimization problems

摘要: Abstract The optimization of truss structures is a complex computing problem with many local minima, while metaheuristics are naturally suited to deal multimodal problems without the need gradient information. Coyote Optimization Algorithm (COA) population-based nature-inspired metaheuristic swarm intelligence field for global that considers social relations coyote proposed single-objective optimization. Unlike most widespread algorithms, its population subdivided in packs and internal influences designed. COA requires few control hyperparameters including number packs, size, maximum generations. In this paper, modified (MCOA) approach based on chaotic sequences generated by Tinkerbell map scatter association probabilities tuning an adaptive procedure updating parameters related condition proposed. It then validated four benchmark planar 52-bar truss, spatial 72-bar 120-bar dome 200 bar-truss discrete design variables focus minimization structure weight under required constraints. Simulation results collected mentioned demonstrate MCOA presented competitive solutions when compared other state-of-the-art algorithms terms quality.