Software Agent Systems

Abstract Agents and multi-agent systems are one of the most fascinating topics in computer science. They attracted and unified not only researchers from nearly all computer science areas but also researchers from other core disciplines such as psychology, sociology, biology, or control engineering. In the meantime, agent-based systems successfully prove their usefulness in many different real-life application areas, especially industrial ones. This is a clear sign that this discipline has become mature. This chapter presents a comprehensive state-of-the-art introduction into advanced software agents and multi-agent systems. Properties and types of agents and multi-agent systems are discussed, which include precise definitions of both. A successful cooperation between agents is only possible if they can communicate in an efficient and semantically meaningful way. Thus, relevant communication strategies are discussed. Agent-based applications can be very powerful, complex systems. Their development can profit a lot from adequate support tools. Different development support options and environments are discussed in some detail. Due to their nature, multi-agent systems are excellent candidates for the realization of comprehensive simulations, especially if the individuality and uniqueness of components of the simulation environment play an important role. The second part of the chapter addresses supporting technologies and concepts. Ontologies, self-organization and emergence, and swarm intelligence and stigmergy are introduced and discussed in some detail.