| In this paper, based on dynamic fuzzy logic, we proposed a mutil-agent coordinate work model in antagonistic environment and non-antagonistic environment in dealing with the mutil-agent system's work mechanism. And out's works include five respects as follows:(1) Based on the principle of yielding, we gave the algorithm to resolve conflicts among agents.(2) Based on the concept of set cover, we gave a generic algorithm to distribute the resource among agents.(3) Based on dynamic fuzzy logic, we proposed a mental state model of agent.(4) By reinforce learning technology, we proposed a on-line learning algorithm of agent.(5) Apply the coordinate model to the workflow manager system for product development.Through the work above, we can describe the natures of the mutil-agent coordinate system and propose the theory base that can help the other researcher build the computer system which can work in coordination friendly. And the work can make up for the inadequate of the exist theory and help find a new way to research agent. |
