Terminal crossover and steering-based particle swarm optimization algorithm with disturbance

Abstract Particle swarm optimization (PSO) is an efficient and simple evolutionary algorithm, which has been successfully applied to solve optimization problems in many real-world fields. Nevertheless, the disadvantage of easy loss of population diversity makes it difficult for particles to jump out of local optimum. The deterioration of population diversity often occurs in the terminal iteration stage. Therefore, to overcome this drawback, terminal crossover and steering-based PSO with distribution(TCSPSO) is proposed in this paper. Firstly, to enhance the diversity of population, a new crossover mechanism is constructed. Meanwhile, in order to make the particle easily jump out of the local optimum, a global disturbance is utilized and the direction of motion of the particle is changed at the later stage. Finally, a nonlinear inertia weight and elastic mechanism are introduced to balance exploration and exploitation better. 34 benchmark functions and two engineering problems are utilized to verify the promising performance of TCSPSO, experimental results and statistical analysis indicate that TCSPSO has competitive performance compared with 15 state-of-the-art algorithms.