Synergistic Model Of Cloud Computing For Power System Based On Mobile-agent

Reliability, economy and safety are the essential elements in power grid developing. These are also factors which have great impact on national economic development. Compared with traditional power grid, smart grid needs to master operation information of power grid accurately, response to information quickly, process tasks intelligently and entirely realize interactive cooperation and resource sharing. The existing centralized power system computing platform has no longer satisfied the reality demand. This condition has become an important barrier that blocks the realization of smart grid. Therefore, the construction of new computing platform for power system has become an important research of scholars and workers.Based on the above analysis, aiming at making up for the lack of information management and service request, Mobile-agent is used in this paper to be regarded as the information interaction mechanism of cloud computing platform for power system, so as to implement synergistic model of cloud computing for power system. This model is layered in accordance with the hierarchical structure of power grid, and realizes resources scheduling and coordinated management via using Mobile-agent. This model can support the computing and shortage capability.Meanwhile, in order to solve the scheduling problem that has the ordering relation in the cloud computing, a scheduling algorithm called Heterogeneous Critical Node Based Duplication is proposed. This algorithm combines list scheduling with duplication. It considers the duplication of tasks before choosing the appropriate processors for them. Moreover, Mobile-agent is brought into scheduling algorithm. Mobile-agent can represent processing capacity of processors and resource available, and it is the intermediary agent between tasks and processors. This algorithm can schedule distributed heterogeneous task which has ordering relation. Overall performance of this scheduling algorithm has been improved by dynamic scheduling and parallel execution.