Abstract: In order to complete task scheduling in grid environment better, and overcome the limitation of the needle current scheduling model, combined with the characteristics of grid task scheduling is a NP ( Non-deterministic Polynomial) problem, this paper proposes a novel grid task scheduling optimization model based on modified SFLA(Shuffled Frog Leaping Algorithm). First of all, grid task completion time is taken as optimization indicators to establish mathematical model, and then leapfrog algorithm is used to find the best scheme for grid task scheduling, and leapfrog algorithm is modified to solve defect and improve the algorithm speed, the grid performance analysis were validated by GridSim. The results show that modified shuffled frog leaping algorithm can complete task scheduling, shorten the task completion time, balance the load of resources, and the performance was obviously superior to the other algorithm.

Key words: task scheduling, leapfrog algorithm, grid computing, resource utilization