Knowledge of the software agent paradigm.

Foundations of software agents: Principles, Architectures, Models. In this course we treat the following topics:

• Agent Architectures
• Multi-Agent Systems
• Agent Communication (Languages)
• Agent Coordination
• Logic-Based Agents

Organization

• Please subscribe to this course in TISS and activate forum participation.
• Register for one group "Anmeldung zur LVA" (I1 to I3).
• Further organizational details (important dates, etc.) will be posted in the corresponding TISS Forum.

Didactic Procedure

• Lecture units alternate with exercise units (3 per term).
• Exercise sheets with deepening exercises have to be solved independently before each exercise unit.
• Necessary knowledge and skills will be presented and illustrated in preceding lecture units.
• During exercise units, solutions to the exercises of the exercise sheets have to be presented on the blackboard; the focus is on presenting the approach, this will not be graded.
• Participating in the test on exercises towards the end of the term, however, is obligatory for a subsequent admission to the final exam.
• The final exam is on the subjects presented during the lecture units.

Addressing software systems that apply software agents first of all calls for the following questions:

• What is a software agent?

Loosely speaking, a software agent is any computer program that provides a service after the request and input of data. E.g., any printer demon, or xbiff are simple software agents. This definition, however, is too simple and further abilities are required, such as to act autonomously, adaptive behavior, or interaction with other agents or the user.
• What are software agents good for?

They are useful in complex, distributed systems whenever services shall be provided by independent, autonomous entities. More complex systems can be built based on the interaction of single agents, each tailored for different tasks and equipped with different capabilities. On the Web, Information Agents play an important role.
• How to build "intelligent" software agents?

There are different approaches (e.g., integrating problem-solving components, special architectures), many of them building on AI methods.

written and oral

siehe http://www.kr.tuwien.ac.at/courses/

Lecture material will be provided via TUWEL.

Basic programming skills (C++, Java) and basic mathematical skills, as taught in the mathematics courses of the bachelor studies, are a necessary requirement. Auxiliary knowledge of knowledge-based systems, as well as of theoretical computer science and logic, is helpful but not necessary.