Research Of Scheduling Algorithm Of Task Based Time Optimization In Mobile Grid

The grid integrates all the resources that come from different departments into one huge "virtual machine" in order to provide an excellent infrastructure and platform for resource sharing and cooperation among distributed computers. So make the best use of the idle resources of network and improve the resource's utilization rate. Mobile grid is a new technology research hotspot sprung up in recent years, which is composed of mobile devices, such as mobile telephone, PDA, Laptop and so on. It dedicates to the sharing and scheduling of mobile resources. Currently, there is litter research focus on mobile devices as resource provider in mobile grid environment, almost all the researches pay attention to use mobile device as an access point to grid system, grid user can ask for grid service by mobile device and receive the task execute result from grid. However, with the performance of the mobile device begin higher and higher, taking the mobile device as mobile grid resource and taking part in grid task is getting too impatient to wait. The thesis will focus on the research about mobile device as a grid service provider.Among all of scheduling algorithm, some work focus on energy consumption, some work focus on deadline constrained scheduling, few work in grid scheduling consider both energy and deadline as constraints. This thesis will not only deal with energy constraining but also deal with the deadline problem in mobile grid. Its primary objective is to minimize the total battery energy in the resource management model as well as to optimize the scheduling time in the task scheduling algorithm.The main research and innovation of this thesis is included in the following areas:Firstly, the idea of Mobile Grid and related infrastructure are proposed and described, which is dedicated to the sharing of mobile resources. Because the centralized manager can result in single trouble and the decentralized manager may consume a lot of battery energy, so these two kinds of management mode are not suitable for mobile grid resource management. This thesis introduced a new kind of mobile resource management——Three Levels Resource Organization Mode, which is suitable for mobile grid environment.Secondly, based on the characteristics of the resources in mobile grid, a novel location management strategy——K steps Pointer Advanced strategy was proposed. On one hand, the strategy can reduce energy consuming caused by updating the resource's location frequently; on the other hand, it can reduce the resource looking up time during task scheduling.Thirdly, based on the limitation of typical Min-Min algorithm, a new algorithm is proposed in this thesis, which named as MG-Min-Min algorithm, the algorithm redefines the Makespan as the sum of the expected completed time and the finding resource's time, and it ignores the network transmission time. The main objective of the algorithm is to minimize the Makespan; the energy optimization will be reflected in the resource management model.Finally, a series of experiments were designed and study the effect on the MG-Min-Min algorithm and the Min-Min algorithm based different parameters, then compare the simulation results with theoretical analysis, as the results show, the MG-Min-Min algorithm turned out to be more suitable for the mobile grid environment because it has considered the mobility of the grid resource and it can get less Makespan than Min-Min algorithm.This thesis is supported by National Natural Science Foundation of China (No:60773211,60970064), the National Science Foundation of HuBei Province under Grant No.2008CDB335, New Century Excellent Talents in university (No:NCET-08-0806), Open Fund of the State Key Laboratory of Software Development Environment(No:SKLSDE-2009KF-2-02), Fok Ying-Tong Education Foundation for Young Teachers in Higher Education Institutions of China (No:121067), NSF of Wuhan Municipality (No:201010621207).