The Study Of Petri Net Model For Grid Task Scheduling Model And Algorithm

Grid computing is the potential platform of the next generation, which will be used in fields of science computing, engineering computing and some other large-scale computing. A good strategy of grid scheduling is needed to allocate resources effectively in order to get the better performance and reduce total execution time and cost in grid computing.Petri net is a kind of graph tool to describe concurrent and asynchronous systems. There are many methods to analyze Petri net, such as reachability graph, reachability tree, incident matrix and so on. Furthermore, its correctness can be proved by mathematics method. Compared with the characteristics of gird, Petri net can be combined with grid computing and study grid as a tool.The paper mostly expatiates on grid task self-scheduling model, the improvement of Min-min algorithm and its modeling, and simulating experiments. The major new work is discussed as follow:Virtual supermarket model is analyzed, which is decomposed into two parts, including resource registering and task self-scheduling. Then the analysis and modeling are provided based on the both parts. And the workflow performance analysis of stochastic Petri net is used to analyze and evaluate the model.Adjustable task scheduling algorithm is presented, which is betterment of Min-min algorithm. A model is built using timed Petri net, and an instance is provided to analyze the executing processes of Min-min algorithm and the better one.A model of Min-min algorithm is presented using Petri net with inhibitor arcs, which is combined with Min-min algorithm and simulates its executing process. Some other kinds of extended Petri net are used to mend the model. Later a more suitable model is provided based on colored Petri net.The algorithm is simulated by GridSim, and the factor on scheduling algorithm is analyzed.