Discrete Particle Swarm Algorithm-based Grid Task Allocation

In recent years, the high-speed networks develop very fast, which makes it possible to integrate the distributed and heterogeneous computing resources into Computational Grid. It provides an ideal platform to solve large-scale computing problems in scientific and engineering area. Because the resources in Grid are distributed, heterogeneous and dynamic, how to schedule tasks in Grid to meet users' requirements is a challenging problem. At the same time, the new intelligent calculation method which is used to resolve complicated optimization problems by simulating natural ecological mechanism, such as genetic algorithm, immunity algorithm, ant colony algorithm and particle swarm algorithm. The scheduling problem with NP-hard optimization became one of the important directions in the research of particle swarm algorithm and ant colony algorithm, since Eberhart and Kennedy presented PSO algorithm in 1995, and Dorigo presented ant algorithm in 1991.A discrete PSO algorithm was proposed and were apply to the problem of gird scheduling in this paper. The algorithm was redefined particle' s position and velocity, and was redesigned the switch rules of position and velocity of particle for grid scheduling .Based on Gridsim ,we designed a grid simulation system , and experiment on the system for the algorithm .Simulation experiment show that it can produce good results , but the algorithm have the local optima in the rear of the algorithm and can' t take full use of the feedback information . On combination of Particle Swarm Optimization algorithm and ant Colony algorithm was proposed for continuous function optimization , the results of PSO were be charge into the start pheromone of ant colony algorithm , the idea of ant search optimize was remodel for continuous function optimization .Numerical experiments showed the good performance of the algorithm . Forther more, the combination algorithm was apply to the scheduling problem of grid task .The discrete PSO algorithm were be the former of combination algorithm , and the result of discrete PSO algorithm were be charge into start pheromone of ant colony algorithm. It solved the blindness search problem of ant colony algorithm. The pheromone update rule of ant algorithm was the reasonable combination on local and all for the performance of ant colony algorithm. Ant colony algorithm were be the rear of combination algorithm that can effectively avoid the local optima of PSO algorithm .The combination algorithm can take full use of PSO algorithm to faster convergence and the positive feedback mechanism of ant colony algorithm...