s Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) are widely used metaheuristics for complex optimization problems, yet both suffer from premature convergence and local optima stagnation. To address these limitations, this study proposes SCGAPSO, a novel hybrid algorithm that integrates GA and PSO via three key innovations: (1) a serial GA-PSO framework that balances global exploration and local exploitation; (2) SINE chaotic inertia weights to diversify search trajectories; (3) a shrinkage factor learning mechanism to refine convergence precision. The performance of SCGAPSO is rigorously evaluated against eight state-of-the-art algorithms (GA, PSO, SA, ACO, ABC, WOA, FOPSO, MLPSO) on CEC2017 benchmark functions and a classic engineering constraint problem—the tension/compression spring design. Experimental results demonstrate that SCGAPSO achieves superior convergence speed, solution accuracy, and robustness. It reduces the average fitness error of complex/mixed CEC2017 functions by 40–60% and achieves 15–30% lower standard deviations on high-dimensional problems. In the highly constrained spring design, SCGAPSO successfully attains the theoretical minimum structural weight (0.01267387). These findings confirm SCGAPSO as a highly robust optimization tool for complex engineering applications requiring both high accuracy and constraint handling.
Wang et al. (Mon,) studied this question.