Study On Service Grid Resourc Scheduling Model Based On Particle Swarm Optimization Algorithm

Grid is a new infrastructure that aims to integrate geographically distributed computing resources. Through this infrastructure, resources in Internet can be fully connected and users can use resources they required without knowing specific details of resources in the infrastructure. The researches on current grid have begun to focus on the field of scientific computing and non-scientific computing from only concerning simple scientific computing. The appearance of the Open Grid Services Architecture shows that the old patterns of resource management can not meet requirements of applications to grid services any more. It has become the trend to provide standardized interfaces of resources in services. Service Grid has combined grid technology and Web services,it represent the direction of development of the grid.Resource scheduling is to make use of scheduling strategy and allocating a reasonable task to different heterogeneous resources in a service grid platform to meet the mandate QoS demand of users. It makes the task implementation of the whole grid system achieve the best performance or the greatest satisfaction of users and realize the goal of shared network resources. In resource scheduling process and sharing entities, different entities may have different QoS(Quality of Service) objectives, which may even contradict each other. This has brought about various possibilities of resource scheduling policies. At present, the hot topics of the grid resource scheduling research focus on coordination among different entities and similar entities within the QoS requirements of the scheduling algorithm. The choice of an efficient resource scheduling solution is the key problem.Particle Swarm Optimization (Particle Swarm Optimization, PSO) is a recently emerging optimization algorithm, with a random start, parallel search and the memory. The concept of the algorithm is simple and easy to understand, it less need to adjust the parameters ,the algorithm simple to achieve ,run fast, and have strong robustness, in other areas of application have been very good.This paper makes a research on the current service grid architecture to discuss the typical grid and resource scheduling systems, and analyses of basic principles of the particle swarm optimization in depth based on the basic concepts of grid features. On this basis, we take into account time and budget for the two QoS constraints which customer needs. We then designed a Grid service resource scheduling mathematical model based on PSO Algorithm and has improved the coding structure of the particles. We use particle swarm algorithm to achieve fast global search to obtain multi-objective optimization solution that satisfies users requirements. Experimental results show that: comparing with other algorithms, the PSO algorithm applied in Grid service resource scheduling has faster response time, a more rational distribution of tasks, an efficient use of grid resources, and have great advantage in solving multi-objective optimization problem.