Concurrent Negotiations for Agent-Based Grid Computing

Since the grid and agent communities both develop concepts and mechanisms for open distributed systems, agent-based grid computing is put forward. However, there are challenges about effective load balancing for grid computing, because of the highly heterogeneous and complex computing environments, even though agents and agent-based grid computing approach are used. To solve the problems, in this paper, a concurrent negotiations model is presented, in which an auction is mapped into a one-to-many negotiation between one seller agent and many buyer agents in service-oriented contexts. Thus, the mechanism and process of concurrent negotiations are studied. An agent negotiates with many other agents and coordinates balance in grid computing resources. The results of exploratory evaluation show that this concurrent negotiations model has a advantage in achieving more and higher utility agreements than other models to optimize computing resources allocation