Improving particle swarm optimization convergence with spread and momentum factors

Particle Swarm Optimization (PSO) is a swarm intelligence search method based on the behavior of birds flocking and fish schooling. It is known for its ability to perform fast computation compared to other evolutionary computational methods like Genetic Algorithms. Several parameter control methods have been developed to make the PSO algorithm faster and more accurate such as linearly decreasing inertia weight (LDIW) and time-varying acceleration coefficients (TVAC). This paper presents an improvement over existing techniques by introducing spread factor and momentum factor into the PSO algorithm. Test results show that the PSO with these two factors produce superior performance and suitable for applications where speed and precision are important.