New Particle Swarm Optimization algorithm for knapsack problem

In this paper it proposes an improved Particle Swarm Optimization (PSO) algorithm for the knapsack problem. The new algorithm is based on the standard PSO algorithm for overcoming the shortcomings that standard PSO traps into local optima easily and has a low convergence accuracy. When the load-bearing quantity of the knapsack is exceeded, the fitness will be sit zero. When the best position of the individual particle is the same with the best position of the population, the particle's position will be reinitialized. The simulation shows that the improved algorithm is simple and effective to solve the small-scale knapsack problem.