| Nowadays Grid has become a hot topic in the world. It can integrate the heterogeneous network resources, coordinate resource sharing in virtual organization that is dynamic and of multi-management mechanism and finally solve the resource sharing problem of large-scale computing. Resource scheduling is the core of the grid system, finishes the mapping from the task set to the resource set. A good resource scheduling strategy is essential to the grid system. It can assign the tasks the right resources as much as possible, greatly improve resource utilization and total cost to make the grid achieve the best performance. Since the resource has the characteristics of heterogeneous, dynamics and autonomy, resource scheduling in grid is more complex and difficult than that in parallel and distributed computing system, it has been proved to be NP-complete. Ant Colony Optimization algorithm has resolved a lot of scheduling problems, for many are NP-hard problem, it has parallelism, strong robustness and is easy to combine with other methods, so we can get good solutions by this way. Therefore, Ant Colony Optimization is very suitable for solving grid resource scheduling. The quality of service (QoS) is also an important measure of the grid, thus, we should consider the users' requirement of QoS in scheduling.First describes the preparation for the design of the grid resource scheduling algorithm, design process and testing about new resource scheduling policy. The methods are discussed as below:(1) Through the study and summary of recent research of grid QoS, propose a grid QoS scheduling model. It focus on three kinds of QoS constraints, gives a detailed description of the various constraints and settings, and introduces the evaluation function to convert the QoS constrains into a utility value to measure.(2)Introduces some grid resource scheduling algorithms, Ant Colony optimization has the character of parallelism is suitable for grid computing, as all the factors that affect resource state can be described by the pheromone, the scheduler can access to predict the outcome simple and fast. In consideration of QoS and Ant Colony Optimization, we propose a QoS-based Ant Colony Optimization scheduling algorithm. This algorithm mainly takes the time, security and reliability into account. In order to consider time and cost, this paper introduces the concept of cost-effective, and make it be the heuristic information of Ant Colony Optimization. On the premise of improve the performance of the grid system, and also taking resource computing ability and cost, to solve the resource scheduling problem better.(3)With the help of the GridSim toolbox, we achieve cost-effective based on QoS-ACO scheduling algorithm for resource scheduling in grid simulation, at last compared with QoS-ACO in total cost value and cost-effective, the algorithm propose in this paper has better result. Due to the extern factors, algorithm proposed in paper has not been applied to the real grid environment, the next step is to apply the algorithm to the real grid environment to test the feasibility and performance. Key words:Grid computing, Resource scheduling, QoS, Ant Colony Optimization, Cost effective. |
PDF Full Text Request