Agents are built based on the BDI framework.

This paper proposes an agent architecture based on BDI model,the architecture integrates main behavior characteristics of deliberative agent and reactive agent,and appling rule database technology to improve the PAC Agent reactive rate.

Owing to the lack of a unified time metric in the BDI model with branching temporal, agents can not make decisions about the goals with time requests and coordination based on such goals.

In this article, we combine the BDI model’s and the non-monotonic logic’s merits, import the conception of possible belief and the operator of belief maintenance to express the interactions between the agents and the dynamic environment, import the conception of the ag.

Belief Desire Intention (BDI) architecture for individual agent is taken to present individual internal mental states and behavior.

The mental state of Agent is constructed on macro-time, and the semantics of concurrent actions are given in micro-time level.

In recent research of modeling agents,most of the work is based on the single mental state,without considering the effect of environment,or are analyzed by just mixing them without considering their internal connection.

Based on ecological principles,we proposed a corresponding eco-thinking model composing eco-thinking of vision,culture,environmental protection and public participation.