Christian Kray, German Research Center for Artificial Intelligence, Germany
This paper presents arguments for the application of a multiagent approach to spatial reasoning, and it is shown how spatial reasoning benefits from an agent-based implementation. In order to achieve this, the basic processes in spatial reasoning are analyzed and mapped onto a multi-agent architecture. Additionally, common problems in the context of real-world applications are identified, and it is shown how these can be addressed using the proposed approach. Furthermore, it is argued that a realization based on a multi-agent system offers advantages in terms of extensibility and flexibility. Finally, an exemplary implementation of a spatial reasoning system within an interactive tourist guide is presented along with its basic mechanisms and a short review of its performance in real world tasks.