A Negotiation Protocol for Optimal Decision Making by Collaborating Agents

Through presented paper, we provide a simple negotiation model in a multi-agent system, where every agent has incomplete set of information about all other agents. Agents in our algorithm collaborate with other agents for optimizing the final solution while preserving their individual privacies. Agents implement the idea of collaboration by sharing their profits with other agents. Therefore an agent prefers existing proposals from other agents to generate its proposals instead of considering proposals solely based on its personal profit. We eliminate the need of a centralizing agent and maintain that minimal information flow is needed among agents while reaching toward an optimal solution. Possible use of our algorithm can be seen in areas where multiple agents are trying to optimize their individual functions through a common solution such as placement of network routers, vehicle routing and job distribution.