Two New Parallel Algorithms Based on QPSO

Abstract

Based on the analysis of classical particle swarm optimization (PSO) algorithm, we adopted Sun’s theory that has the behavior of quantum particle swarm optimization (QPSO) algorithm, by analyzing the algorithm natural parallelism and combined with parallel computer high-speed parallelism, we put forward a new parallel with the behavior of quantum particle swarm optimization (PQPSO) algorithm. On this basis, introduced the island model, relative to the fine-grained has two quantum behavior of particle swarm,m optimization algorithm, the proposed two kinds of coarse-grained parallel based on multiple populations has the behavior of quantum particle swarm optimization (QPSO) algorithm. Finally under the environment of MPI parallel machine using benchmark functions to do the numerical test, and a comparative analysis with other optimization algorithms. Results show that based on the global optimal value is superior to the exchange of data based on local optimum values of exchange, but in the comparison of time is just the opposite.