Font Size: a A A

The Research Of Grid Scheduling Based On Ant Colony And Genetic Algorithm

Posted on:2010-02-13Degree:MasterType:Thesis
Country:ChinaCandidate:J LiuFull Text:PDF
GTID:2178360275479662Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Grid which gathers computation resources,storage resources,knowledge resource, communication resource,and information resources etc.together from all over the world, aims at serving public and realizing resources sharing and cooperative working.Grid technologies,with its extensive developing perspective and commercial value,have been a hot topic both at home and abroad research.However,there remain quite a lot of key technical issues waiting to be resolved in Grid,among which,Grid task scheduling is especially outstanding.The dynamic change,geographical dispersion and heterogeneous systems of Grid complex the Grid task scheduling.Therefore,it is necessary to work out a dispatching scheme adaptable in Grid computation environment.The scheme could improve the efficiency of user's task completion so that the service quality could be satisfactory to the users.Bionic optimization algorithm applied in Grid task scheduling becomes a sharp tool to solve the problem of Grid task scheduling.As one of the bionic optimization algorithm, Ant colony algorithm has been applied in Grid task scheduling by more and more scholars with a good effect for its dynamic and self-similarity with the theory of Grid task scheduling.But they do not take parameter optimization into account and value inspiring information parametersα,the expectations factorβ,information volatile factorρ,pheromone strength Q according to previous experiment.In fact,the convergence performance and computational efficiency of ant colony are sensitive to it.An ant system based exploration-exploitation for reinforcement learning.In this paper,we integrate genetic algorithm into ant colony algorithm and form ant colony genetic algorithm.This algorithm uses fast global search randomly of genetic algorithm and explores optimization of four parametersα,β,ρ,q to achieve a more reasonable scheduling.
Keywords/Search Tags:Parameters grouping, Ant colony, Genetic algorithm, Ant colony and genetic algorithm, GridSim
PDF Full Text Request
Related items