From the Specification of Multiagent Systems by Statecharts to Their Formal Analysis by Model Checking: Towards Safety-Critical Applications

In order to design and implement multiagent systems, the specification method should be as expressive and comprehensive as possible. Statecharts, which are capable of describing dynamic systems and are widely accepted in the computer science community, are applied and investigated for this objective. In this paper, multiagent systems are studied in the domain of robotic soccer, where the behavior of agents including collaboration is specified by means of UML statecharts [7,8]. This method is also applicable to industrial software applications. For example, a network application can be designed and specified by means of the same method (see [1]). The approach enables not only standardized design of multiagent systems, but also almost automatic translation of the specification into a running implementation. As a natural extension of this methodology for designing multiagent systems, formal system analysis and verification should be possible, in order to investigate important system properties. In this paper, we will show how specifications of multiagent systems based on UML statecharts can be verified, by employing model checking techniques. Hence, the proposed specification technique can be used for both, automated multiagent system implementation and analysis.