Abstract:

Using the Particle Swarm Optimization (PSO) algorithm to solve the problem of virtual network embedding can reduce the consumption of network resource, but it also brings premature convergence.An improved PSO algorithm is proposed,which adds random factors,operates along the original direction, and changes the introduction  of history factors to the search process.The proposal not only keeps the instruction of history factors to the search process but also increases the search range,thus relieving, the premature convergence problems to a certain extent.The experimental results demonstrate that the improved PSO algorithm can be applied to virtual network mapping and effectively reduce resource consumption in comparison with the original PSO algorithm.

Key words: virtual network mapping;premature convergence;particle swarm optimization (PSO)