Research Of Grid Task Scheduling Algorithm Based On Ant Colony Algorithm

This paper introduces the contents of the grid and related technologies, on this base, do a research on the core issues:task scheduling technique, which affects the service quality and performance of the grid. Extend the field of application of the ant colony algorithm, a method using an improved ant colony algorithm to solve the problem of grid task scheduling is proposed. Two points are improved specifically:First, introduce the quantitative indicator load balancing degree F which reflects grid load balancing situation good or bad. The value of F is smaller, load balancing of grid is thought to be better, ants would secrete more pheromones when passed relevant path. On the contrary, the value of F is larger, the amount of pheromones is less. Second, adjust the pheromone evaporation factor adaptively, update and retain optimal solution at the end of each cycle to avoid the situation that algorithm appears premature convergence or falls into local optimum.According to the theory above, build a mathematical model of the algorithm and do the simulation. Through experimental comparison in improved ant colony algorithm, basic ant colony algorithm and min-min scheduling algorithm, we found the algorithm proposed can effectively shorten the grid task completion time span and improve the load balancing of grid.