After completion of the course, the student can explain basic aspects of multi agent systems such as agent architectures, interaction, intelligence, negotiation and communication. The student will gain overall knowledge of multi-agent concepts and master a specific topic of Multi Agent Systems in depth. The student will be familiar with seminal and novel scientific studies in the multi-agent domain. Through discussion of sample applications, the student will understand opportunities and barriers of implementing multi-agent techniques in practice. The student will gain experience in multi-agent systems research by choosing an appropriate research question, agent design methodology and implementation platform. The student will be able to design and implement a basic multi-agent system using design science methods and learn how to report and reflect on multi-agent systems. Working in a small group, the student will produce and present a scientific paper on the project in English.
|
|
An agent is a software or hardware module that is able to pursue, in an autonomous and rational way, one or more goals given to it by a user or another software system. In the course Multi-Agent Systems, we study computer systems that consists of such agents. In multi-agent systems, individual goals of agents are often conflicting. Therefore, agents have to co-operate to solve these conflicts and accomplish their goals. Co-operation may also be necessary if one individual lacks sufficient resources to accomplish its own goals. In a multi-agent system, co-operation is never built in a-priori by a central system designer. Instead, agents have to organize co-operation themselves, where each agent constantly takes its own self-interest as a starting point. Topics covered in the course are Agent Architectures, Interaction, Negotiation and Communication, some degree of ‘intelligence’, applications of agent systems, recent developments such as learning agents/bots.
|
|