A negotiation model through multi-item auction in multi-agent system

Agents often work together and achieve their common and individual goal under the environment where there are limited or rare resources. How to allocate these resources efficiently is critical in proceeding of cooperation and competition in multi-agent system. This paper proposes and investigates an approach to solve the problem by introducing auction theory from economics game. An agent-based multi-item auction model is studied as negotiation method so as to get some precious resources for buyer agents and allocate them for vendor agent. Some properties and impacts of agents' parameters are shown through simulating experience such as buying power, item number required by buyer agent, number of buyer agents, rules of auction and so on. All this is key and beneficial for the future mechanisms design both in agent-based software engineering and various distributed system, especially Internet-based commerce applications.