A Novel Probability Binary Particle Swarm Optimization Algorithm and Its Application

Particle swarm optimization (PSO), an intelligent optimization algorithm inspired by the flocking behavior of birds, has been shown to perform well and widely used to solve the continuous problem. But the traditional PSO and most of its variants are developed for optimization problems in continuous space, which are not able to solve the binary combinational optimization problem. To tackle this problem, Kennedy extended the PSO and proposed a discrete binary PSO. But its performance is not ideal and just few further works were conducted based on it. In this paper, we propose a novel probability binary particle swarm optimization (PBPSO) algorithm for discrete binary optimization problems. In PBPSO, a novel updating strategy is adopted to update the swarm and search the global solution, which further simplify the computations and improve the optimization ability. To investigate the performance of the proposed PBPSO, the multidimensional knapsack problems are used as the test benchmarks. The experimental results demonstrate that PBPSO has a better performance for solving the multidimensional knapsack problem in terms of convergent speed and global search ability.