A Comparative Study on Different Parameter Factors and Velocity Clamping for Particle Swarm Optimisation

<jats:title>Abstract</jats:title>
<jats:p>Optimisation is an interesting topic because various optimum results can be obtained using limited resources. This is the reason that researchers are interested in further investigating the possibility of optimisation by testing numerous optimisation methods. As the current research trends are preferred to solve the problems using Artificial Intelligence (AI) methods, hence more nature-inspired optimisation methods are developed. Particles Swarm Optimisation (PSO) is one of the well-known optimisation methods in Swarm Intelligence (SI). It can be used to solve many problems. However, PSO is easily trapped in local optimum and encountered premature convergence problem, which indirectly affects its performance. Besides that, its performance with different parameter factors and velocity clamping is always not considered by researchers. Therefore, the objective of this research is to compare the performance of PSO with different parameter factors and velocity clamping using average fitness values. 10 different types of PSO models are applied and tested using 10 benchmark functions. Based on the experimental results, PSO10 with constriction factor and velocity clamping is proved to be the best PSO model because it can obtain 10 out of 10 minimum average fitness values. The outcomes can assist the researchers while applying the method to their research.</jats:p>