A Particle Swarm Algorithm for Complex Quantised Problem Spaces

The particle swarm algorithm has shown ability to optimize in continuous problem spaces, although it can struggle in problem spaces containing multiple optima. A variant, called Waves of Swarm particles (WoSP), has been shown to be able to handle problem spaces containing multiple optima by sequentially exploring these optima. In this paper the WoSP algorithm is adapted to suit complex quantised problem spaces and applied to a highly constrained problem with many constraint-violating solutions but few constraint-satisfying solutions. The performance obtained is remarkably good and reflects the power of the WoSP algorithm that combines the search ability of particle swarm with that of evolution.