Research On Decision Support System At Grid Environment

Grid computing can provide more resource, function and interactivity for grid users. We should still put focus on the resource utilization with regard to such a distributed system of large-scale grid. In the grid, the existing nodes may quit for all sorts of reasons at any time, and there will also be many new nodes involved, which makes the grid resource discovery mechanism should return one or more resource matching positions. So the resource discovery is a very important problem. Today, the other research about grid has been matured, under this circumstance, the task of resource discovery will be more and more.In order to judge whether the node applies sub-grid, we need to consider the multi-attributes in the process of using resource discovery mechanism to return new nodes, and to make those join the grid. The grid resources process various attributes, and the Comprehensive Evaluation Principle is to analyse evaluation target in the case of multiple indices. So using the Comprehensive Evaluation Principle of judgement is a good choice.In the field of grid computing, stress of grid resource should be put on providing services, but it will be charged in the future. So the increasing of network node implications the change of economic cost in direct ratio. This paper on the basis of layered structure, proposes a discovery and access strategy in view of comprehensive index method, which aims at resource cost in the field of grid. This strategy can comprehensively consider two factors----execution cost and execution time, and then comprehensively evaluate the resource who wants to join the grid actively. On the other hand, the entities in the grid are changeable, so it will cause node failure problem inevitably. To ensure the reliability of the node in the field of information grid and solve the security problem on the dynamic grid environment, comprehensive index method was used to build a comprehensive evaluation model, which was for the direct trust and indirect trust between grid nodes, and the evaluation density of the node in the unit time. This model can extract trust information from the interaction between the entities in the grid, and then evaluate one node's overall trust degree synthetically. The entities in the grid were controlled interactively by setting system threshold.This paper, under the premise of that network model discovery resource smoothly, used comprehensive evaluation method to estimate resources in multi-attributes, such a method builds correponding sub-grid that can optimize the cost and function. The Simulation shows this model can improve grid mission success rates, and reduce the influence of the grid uncertainty.