Research Of Task Scheduling Algorithm With Energy Optimization Based On GA In Grid

Grid computing is the latest emerging research which has a high academic value and application value.Grid computing involved a lot of key technologies such as: resource description and discovery,resource management and organization, computing task scheduling,performance evaluation,communication protocol specification,security system,fault tolerance mechanism and so on.As to grid task scheduling,the problems are faced with resource-constrained.A mobile ad hoc grid is a heterogeneous computing and communication system that allows a group of mobile devices to accomplish a mission,so energy management is a major concern in ad hoc grids.Task scheduling based on grid energy problem here focuses on statically assigning resources in an ad hoc grid to an application composed of communicating subtasks.The goal of the allocation is to minimize the average percentage of energy consumed in the ad hoc grid.The main research topics are grid task scheduling and energy optimization.Grid task scheduling consider the first complete time,that is,makespan,while energy optimization considered how to statically assign resources to inter-corresponding sub-task requests in the ad hoc grid.The goal of allocation is to minimize the average energy consumption when this request is implemented in ad hoc Grid.Among them, the key issue is the allocation of energy resources.Its purpose is to consider the smallest makespan and the least energy consumption.The main research and innovation of this thesis is in the following areas:Firstly,research and analysis of typical grid task scheduling algorithm.Typical grid task scheduling algorithms include Min-min and Max-min algorithm,taboo algorithm,and genetic algorithm and so on.These algorithms have some advantages, but for grid task scheduling,which is an NP problem,the genetic algorithm simulate evolution is the best choice.Second,a method of optimizing energy of grid is put forward on basis of the traditional grid energy model.Grid resource can be divided into three categories: computing resources,storage resources,information resources(including software resources,knowledge resources,etc.).According to this classification can receive generalized energy function.A variety of grid energy models include energy model of ad hoc grid static resource,DVS energy aware scheduling model,energy model of task allocation for embedded systems.A static method of optimizing energy of grid is put forward based on the above research.Thirdly,Grid task scheduling algorithm based on genetic algorithm is put forward according to energy of grid and makespan.In this thesis,grid task scheduling algorithm based on genetic algorithm is put forward,which combines the energy grid and makespan of QoS bound of user.The algorithm meets the need of minimum of makespan and energy consumption of the grid,at the same time the algorithm saves maximally energy consumption of grid in the most priority scheduling of grid task.The key of this thesis is algorithm research.First,feasibility of the algorithm is introduced;theoretically mathematical analysis and optimization model are put forward about the algorithm.Simulation test is made in grid simulation environment about the proposed algorithm.The result showed simulation and theoretical analysis are consistent.Finally,typical grid scheduling simulation tools and its characteristics are analyzed,SimGrid is chosen as the simulation platform in this research,the simulation processes and method is introduced,then the proposed algorithm is simulated in SimGrid,and simulation results and theoretical analysis are compared and analyzed.