A Task Scheduling Strategy In Computational Grid Based On Genetic Algorithms And Ant Algorithms

Grid computing is a kind of perfect computing technology, which will be used in the fields of science computing, engineering computing and some other large-scale computing. As a centrum of entire grid computing, grid core service, which connects grid understratum groupwares to top groupwares and takes charge of entire grid system in order to ensure grid system work effectively, is an important part of grid computing. It is important to research grid core service technology. Task scheduling technology is a part of grid core service technology.When grid resources are required by lots of tasks, the system can optimize the resources only by scheduling the tasks reasonably. By harmonizing and distributing the grid resources efficiently, an advanced task scheduling strategy can reduce total run time and total expense greatly and bring an optimal performance, which makes the task scheduling strategy a key technology to grid computing.At present, a number of grid tasks scheduling algorithms, such as genetic algorithm(GA), ant colony algorithm(ACA), simulated annealing algorithm and sufferage algorithm etc, have advantages and disadvantages obviously, which can not finish optimizing the task scheduling by itself and a strategy that combines GA and ACA is not brought forward. By learning from the former two algorithm'strong points to offset their weakness, the paper analyses the advantages and disadvantages of the two algorithms at different stage particularly and present a new scheduling strategy based our designing model. The innovation of this strategy is to form the initializing key quickly at former stage by utilizing GA until meets the terminating condition, and transforms the key into pheromone needed by ACA, subsequently makes use of the character of ACA to find the optimal result rapidly. The aim of this new strategy is to schedule resources more reasonably and improve capability of grid computing by this means. In order to evaluate the performance, we design a simulating program to validate it after finishing the SimGrid study.The thesis mainly expatiates the preparation, designing process, implementation and performance simulation of the strategy.