An improved hybrid particle swarm optimization with dependent random coefficients for global optimization

Abstract

In this paper, an improved hybrid particle swarm optimization (IHPSO) was proposed by using the learning strategies framework of the particle swarm optimization (PSO), and adapting the gravitational search algorithm (GSA) into the PSO. To be specific, the IHPSO adopts three learning strategies, namely dependent random coefficients, fixed iteration interval cycle, and adaptive evolution stagnation cycle. The particle first enters into the PSO stage and updates its velocity based on the first strategy to enhance the exploration ability. Particles that fail to improve their fitness then enter into the GSA operators in terms of the latter two strategies to decrease the computational cost in the hybridization. To evaluate the effectiveness and feasibility of the IHPSO, the simulations were performed on various test functions. Results reveal that the IHPSO exhibits superior performance in terms of accuracy, reliability and efficiency compared to PSO, GSA and other recently developed hybrid variants.