Fitness proportionate selection based binary particle swarm optimization

Particle Swarm Optimization(PSO) has shown its advantages not only in dealing with continous optimization problems, but also in dealing with discrete optimization problems. Binary Particle Swarm Optimization(BPSO), the discrete version of PSO, has been widely applied to many areas. Although there are some variations aiming to improve BPSO's performance, none of them has been proven to be a promissing alternative. In this paper, we propose a novel binary particle swarm optimization called Fitness Proportionate Selection Based Binary Particle Swarm Optimization(FPSBPSO). We test FPSBPSO's performance in function optimization problems and multidimension knapsack problems. Experimental results show that FPSBPSO can find better optima than BPSO and a variation of BPSO.