Job Scheduling of Savant for Grid Computing on RFID EPC Network

Yl Pan,Yc Lee,F Wu

DOI: https://doi.org/10.1109/scc.2005.59

2005-01-01

Abstract:In the Grid computing environment, there are many important issues, including information service, information security, resource management, routing, fault tolerance, and so on. Job scheduling is a major problem since it is a fundamental and crucial step in achieving high performance. The job scheduling problem has been treated as a combinatorial-optimization problem. Scheduling in a Grid environment can be seen as an extension to the scheduling problem on local parallel systems. In the research, we focus on applying the technologies of RFID and Grid computing to the architecture of the EPC network. There are many applications needing computing power of the Grid to deal with the huge quantity of incoming EPC data. According to the architecture of the EPC network and the environment of Grid computing, they have several similar characteristics. Therefore, we propose a new algorithm that modifies the traditional GA and integrates SA. Since the processes of GA and SA keep no memory, some problems may be visited again. In order to overcome this drawback, we design a learning scheme to remember visited statues to reduce the probability of the re-visiting improvement the performance in the search space. It can help to find the optimal or near-optimal scheduling efficiently and avoid the resource deadlock. Furthermore, our proposed algorithm, HGASA - with learning scheme, also considers about several properties of grid computing environment and EPC network, such as heterogeneity, dynamic adaptation and the independent relationship of jobs.