Cooperative Problem Solving Process Based on MAS_NP Model

Abstract

In order to enable the individual agents to be aware of the both macro constraints from agent social and instructions from agent's owner, a norm and policy extended agent BDI model (called NPCD-Agent) has been proposed, by integrating norms and policies into traditional agent BDI model. On the basis of NPCD-Agent model, this paper proposes a semantic model of multi-agent system (called MASNP), where cooperative problem solving (CPS) process is formally described. On the one hand, at the social level, the social constraint of agents can be specified by means of norms defined by the social administrators. And on the other hand, agents are instructed by polices presented by their owners. A case study shows that macro social constraint and instruction from agent's owner are both considered in CPS process, which makes the CPS process more reasonable.