Tasks scheduling is a key problem in multi-agent system, traditional tasks scheduling methods can't be applied to new application areas of the MAS such as emergency system. In order to apply agent method to these ...
详细信息
ISBN:
(纸本)9780769536187
Tasks scheduling is a key problem in multi-agent system, traditional tasks scheduling methods can't be applied to new application areas of the MAS such as emergency system. In order to apply agent method to these new areas, this paper proposes a new variant of task scheduling method considering utility. Based on our MAS model, we construct a flow network according to instance of task scheduling problem and add utility to it. We show that minimum cost flow algorithm can be use to task scheduling problem to maximize the utility of the system. Moreover, we use a mathematical method to ensure tasks are schedulable when maximizing the utility. Finally, a conclusion which can be used to determine whether there exists an efficient task scheduling is gained.
In multicomputers that utilize circuit switching or wormhole routing, communication overhead depends largely on link contention-the variation due to distance between nodes is negligible. This has a major impact on the...
详细信息
In multicomputers that utilize circuit switching or wormhole routing, communication overhead depends largely on link contention-the variation due to distance between nodes is negligible. This has a major impact on the load balancing problem. In this case there are some nodes with excess load (sources) and others with deficit load (sinks) and it is required to find a matching of sources to sinks that avoids contention. The problem is made complex by the hardwired routing on currently available machines: the user can control only which nodes communicate but not how the messages are routed. Network flow models of message flow in the mesh and the hypercube have been developed to solve this problem. The crucial property of these models is the correspondence between minimumcostflows and correctly routed messages. To solve a given load balancing problem, a minimum cost flow algorithm is applied to the network. This permits us to determine efficiently a maximum contention free matching of sources to sinks which, in turn, tells us how much of the given imbalance can be eliminated without contention.
暂无评论