Reasoning Based on Rough Set Among Multi-Agents Under the Environment of ABDIDSS

Models of agent and ABDIDSS (ABDIDSS--distributed intelligence decision support system based on multi-agents) are defined in this paper. Approximation spaces of the agent reasoning in ABDIDSS are built up. Rough set on approximation spaces is defined. Reasoning in ABDIDSS may be realized through the means of establishing decision trees. Multi – agents comprehending to definable sets, synthesis operation of sets and solution of lower and upper approximations of sets on approximation space are included on decision trees. Therefore methods of multi-agents comprehending to definable sets, synthesis operation of sets and solution of lower and upper approximations of sets on approximation space are defined under the environment of ABDIDSS. They have important theoretical meaning on approximation reasoning based on Rough set. It is shown with real example that it is feasible to realize problem to be solved cooperatively among multi-agents in the last.