A generalized intelligent agent model and its application in the domain of transportation research

Intelligent agent technology is the articulation of human decision-making behavior in the form of a computer program. From software engineering and from e-commercial perspectives intelligent agents are expected to become a big success in the future as they provide intelligent assistance in situations in which interfaces for the direct manipulation of software or of information are insufficient. It can be assumed that most future software systems contain components which can be traced back to agent technology. However, from the scientific perspective today's achievements are small and the goal to design agents that can act in open and distributed environments in which loosely coupled components interact is a key challenge for an overwhelming future of intelligent agent technology. This thesis discusses fundamental issues of intelligent agents and it provides a general agent architecture emphasizing aspects of perception, interpretation of natural language, learning and creativity. Agents built upon this architecture are equipped to handle unknown, open and distributed environments. Particular attention is drawn to options how the discipline of transportation research can benefit form the potential of intelligent agent technology. To illustrate this, a microscopic multi-agent based road traffic simulation environment, based on the general agent architecture, is described. The thesis concludes with a vision of future in-vehicle route guidance systems utilizing personalized and interacting intelligent agents