Particle swarm optimization (PSO) is a kind of evolutionary algorithm to find optimal (or near optimal) solutions for numerical and qualitative problems. Recently, a new variation on the traditional PSO algorithm, called cooperative particle swarm optimization (CPSO), has been proposed, employing cooperative behavior to significantly improve the performance of the original algorithm. However, a standard CPSO is focused only on continuous problems. In this paper, we present a new approach based on the CPSO to solve combination optimization problems by introducing dynamic splitting schemes. Reverse operation and simulated annealing techniques are further used to prevent the algorithm from being trapped in local minima. Finally, Traveling salesman problem (TSP) is applied to show the effectiveness of the proposed PSO.