A novel orthogonal PSO algorithm based on orthogonal diagonalization

摘要: Abstract One of the major drawbacks global particle swarm optimization (GPSO) algorithm is zigzagging direction search that leads to premature convergence by falling into local minima. In this paper, a new named orthogonal PSO (OPSO) proposed not only alleviates associated problems in GPSO but also achieves better performance. OPSO algorithm, m particles are divided two groups: one active group best personal experience d and passive remaining (m ‒ d) particles. The purpose creating groups enhance diversity swarm's population. each iteration, undergo an diagonalization process updated such way their position vectors orthogonally diagonalized. as contribution finding correct significant. using guide, thus avoiding conflict between guides occurs algorithm. We tested with thirty unimodal multimodal high-dimensional benchmark functions compared its performance several competing evolutionary techniques. With extensive simulated experiments, we have shown superiority the proposed terms convergence, accuracy, consistency, robustness reliability over other algorithms. found be successful achieving optimal solution all functions.