1: \begin{abstract}

2: 

3: An information agent is viewed as a deductive database consisting

4: of 3 parts:

5: \begin{itemize}

6: \item an observation database containing the facts the agent has

7: observed or sensed from its surrounding environment.

8: \item an input database containing the information the agent

9: has obtained from other agents

10: \item an intensional database which is a set of rules for

11: computing derived information from the information stored in the

12: observation and input databases.

13: \end{itemize}

14: 

15: Stabilization of a system of information agents represents a

16: capability of the agents to eventually get correct information

17: about their surrounding  despite unpredictable environment changes

18: and the incapability of many agents to sense such changes causing

19: them to have temporary incorrect information. We argue that the

20: stabilization of a system of cooperative information agents could

21: be understood as the convergence of the behavior of the whole

22: system toward the behavior of a ``superagent'', who has the

23: sensing and computing capabilities of all agents combined. We show

24: that unfortunately, stabilization is not guaranteed in general,

25: even if the agents are fully cooperative and do not hide any

26: information from each other. We give sufficient conditions for

27: stabilization. We discuss the consequences of our results.

28: 

29: \end{abstract}

30: