A Research Of Multi-object Grid Task Scheduling Based On Niche-genetic-tabu Search Algorithms

With the development of grid technology and in-depth study , geographically dispersed and heterogeneous resources can be organized into a virtual grid supercomputer. Grid task scheduling is how to carry out the most effective management and how to effectively complete all kinds of computing tasks by use of giant Grid sources. Task scheduling is an important part of grid computing. It has always been an active research area in international and domestic academic circles. The thesis is to develop a fusion niche genetic tabu search algorithm and its application in multi-objective grid task scheduling application research.The problems of grid task scheduling are analyzed through summarizing research content and significance of grid task scheduling and research actuality.Genetic algorithm and tabu search algorithm are powerful tools to solve the complicated large-scale optimization problems. To deal with the prematurity and low convergence speed when the genetic algorithm being used for global optimization and Tabu search depend on its initial solution strongly , through comprehensive contrast and comparison between the above two algorithm, a hybrid optimization algorithm was introduced to improve the local search ability of Genetic algorithm. In this algorithm, in order to speed up convergence speed and get satisfied results, Tabu search algorithm was used for local search, Genetic algorithm was applied for global search. Meanwhile niche was imported to control prematurity and to avoid converging to local optimum. The test results show that both calculating speed and output are improved, so it is a fast and effective algorithm.Response in a dynamic, complex grid systems, resource failure is very frequent, affecting service quality and efficiency of grid computing problems, the above algorithm( NGATS) is applied to multi-objective grid task scheduling, integration of the formation of a new niche Multi-objective genetic taboo Grid Task Scheduling Algorithm (NGATS-MOGTS), the survival of the task and task completion time (Makespan) combined give an adjustable multi-objective integration of the utility function. Performing an experiment based on Gridsim to emulating the algorithm NGATS-MOGTS. The simulation results show that the scheduling algorithm can trade off these two objectives. So it can be applied in the complex grid computing environment well.