Improved particle swarm optimization based on boundary condition and its application for task scheduling on grid
To solve the problem of slow search speed and premature convergence of particle swarm optimization (PSO) algorithm in the late period of evolution, an improved PSO algorithm with boundary condition is proposed, where Search Space-zoomed factor and Attractor are employed to deal with the particles flying outside search space by simple and effective calculation. And grid task scheduling based on the proposed PSO algorithm was designed to improve the grid performance and implemented in simulation environment of dynamic and heterogeneous grid systems. The comparative experiment results show the proposed PSO algorithm in this paper is better to gain the shortest completion time of application, reasonably allocate the resources in grid system, and efficiently balance workload simultaneously than other experienced algorithms from the literatures.
Author's Name: Chi, Y., Sun, F., Wang, W., Wang, S., Yu, C.