Abstract
A novel orthogonal multi-swarm cooperative particle swarm optimization (PSO) algorithm with a particle trajectory knowledge base is presented in this paper. Different from the traditional PSO algorithms and other variants of PSO, the proposed orthogonal multi-swarm cooperative PSO algorithm not only introduces an orthogonal initialization mechanism and a particle trajectory knowledge base for multi-dimensional optimization problems, but also conceives a new adaptive cooperation mechanism to accomplish the information interaction among swarms and particles. Experiments are conducted on a set of benchmark functions, and the results show its better performance compared with traditional PSO algorithm in aspects of convergence, computational efficiency and avoiding premature convergence.
Highlights
The Particle Swarm Optimization (PSO) algorithm is a stochastic, population-based evolutionary optimization technique and was conceived in 1995 by Kennedy and Eberhart [1,2]
It is beneficial to maintain a frequent information interaction among those swarms dispersed process progresses, it is more beneficial to decrease the frequency of information interaction in swarm level and increase the frequency of information interaction in particle level, so that the swarms could effectively obtain an optimum in a local space
The information interaction mechanism in the swarm level in our MCPSO‐K should in the search space in the beginning of the optimization
Summary
The Particle Swarm Optimization (PSO) algorithm is a stochastic, population-based evolutionary optimization technique and was conceived in 1995 by Kennedy and Eberhart [1,2]. Each particle in the PSO algorithm represents a potential solution and aims to search a specific search space with a velocity which is dynamically adjusted according to a set of novel velocity and position updating rules reflecting a particle’s own and its companions’ historical behaviors. The PSO algorithm has been studied and applied to solving complex mathematical problems and real engineering optimization problems [3,4,5,6]. Similar to most stochastic algorithms, the performance of PSO algorithm will suffer from a significant degradation when it is applied to multi-dimensional or tightly coupled problems. To improve the performance of the traditional PSO algorithm and overcome the disadvantages of traditional
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have