Multi-Agent-Based Negotiation Algorithm Oriented to Multi-Ship Collision Avoidance

Multi-ship collision avoidance is always a hot and difficult issue in marine field. Thereinto, the accurate expression for shipping behavior and intent is key problem, as well as negotiation for collision avoidance. In this paper, A model of decision system for multi-ship collision avoidance is proposed based on Multi Agent System (MAS) on the speciality of opening, distribution, multi-source heterogeneous data, multi-demands and complex function. Hierarchical structure is adapted in the model. The advantage of the model is that communication relation is optimized, resource is retrenched, and manages and control for agents is convenient. In view of problems that in collision avoidance situation, negotiation is based on the optimal of self-benefit, whereas incomplete information for other agents, this paper proposes a Negotiation Algorithm based on debate strategy among MAS, and analyses the complexity of the algorithms.