Based On The Expectation Trust Benefit Function Of The Grid Job Scheduling Algorithm Research And Simulation

Grid is an important information technology in world in recent years. It makes users can share many different resources in different areas. Task schedule is an important part of this field. According to the task condition, schedule allocates appropriate resources to different tasks by rational strategies. The different constructs and dynamic of the grid system and applications in the grid require varies of resources bring task schedule more complex.Therefore, in this thesis, bases on the analysis of Grid scheduling, it introduces the application of the trust management on Grid. Then, we use Dempster-Shater theory describer the trust function, the likelihood function of Grid. Main innovation project is the based expect credit profit drive task schedule method. The expect profit function is the forecast gain of one task executes on Grid node. This thesis proposes the expectation trust benefit driven Grid job schedule arithmetic. The value of the expectation trust benefit function is prediction benefit of a certain job which executes in Grid. Simulation proves that expectation trust benefit driven arithmetic is better than traditional min-min arithmetic in benefitSimulation proves that expectation benefit driven arithmetic is better than traditional min-min arithmetic in benefit and it also has good performance in benefit.