Optimizing Convergence and Diversity: A Synergistic Approach using Genetic Algorithm and Particle Swarm Optimization

Abstract

PSO (Particle Swarm Optimization) is a widely used optimization technique. One notable variation of PSO is Auto Improved Particle Swarm Optimization (AIPSO). Due to its simplicity, PSO is widely used in many applications; nevertheless, although AIPSO converges quickly, it suffers from a noticeable stagnation problem. The Auto Improved Particle Swarm Optimization strategically upgrades itself by utilizing Genetic Algorithm (GA). This combination adds necessary variation, offsets the early convergence that is present in AIPSO, and successfully tackles the issue of stagnation. Combining AIPSO with Genetic Algorithm improves convergence and gets rid of stagnation problems. By taking use of the genetic algorithm's variety provision, the hybrid GA-AIPSO technique improves algorithm performance overall by preventing premature convergence to AIPSO-generated solutions. A detailed comparison with the most advanced algorithms, including JADE, GA-DE, and PGHA, confirms that the suggested hybrid GA-AIPSO is more successful. The algorithm's ability to break through stagnation and accelerate convergence rates in optimization problems is clearly demonstrated by the results.