A Noval Reduced Particle Swarm Optimization With Improved Learning Strategy and Crossover Operator

Abstract

In terms of balancing the exploration and exploitation capabilities of the PSO method in order to increase its resilience, this work provides a unique particle swarm optimization with enhanced learning techniques and a crossover operator (LSCPSO). Each particle is updated depending on the simplified equations in the first stage. The proposed LSCPSO method then employs a self-learning technique in which each particle (personal best) learns from k better particles in the current population. Then, a crossover step is introduced to the algorithm in the subsequent stage. After taking the k global best (gbest particle), the crossover is performed. This method strengthens the LSCPSO algorithm's capacity for social learning and global exploration. In subsequent trials, the performance of the LSCPSO algorithm is compared to that of five sample PSO variations. The benchmark function test results show that the proposed ILSPSO algorithm has much better overall performance than the other PSO variations that were looked at.