Multiagent Systems

Content:
This course provides insights in models and methods of the field of multiagent systems, a new paradigm for modeling, analysing, and predicting complex sociotechnical systems. We are interested in appropriate software architectures and models for the expression of cooperation and competition between intelligent software systems, as well as between humans and these software systems.

Learning Objectives:
Students are familiar with models and architectures of intelligent autonomous agents. They understand modeling levels of socio-technical systems and their realization by means of models and mechanisms of multi-agent systems. They understand the main implications of the rationality vs. cooperation assumption. They know essential programming languages for the implementation of multi-agent systems (in particular logical programming, concurrent models and the BDI paradigm) and can use them to design and implement smaller multi-agent systems. After completing the module, students will have knowledge of the most important theoretical foundations of multi-agent systems, in particular decision models with game-theoretical concepts. They will be able to consider and apply the acquired skills in the development of distributed cooperative systems and use them for analysis.

Link to the repositories: TBD

Extent: 6 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K01 Introduction to multiagent systems

Content:
Students are given an introduction to multi-agent systems as well as socio-technical systems and autonomous systems.

Learning Objectives:
After completion students understand modeling levels of socio-technical systems and their realization using models and mechanisms of multi-agent systems. They understand the properties of the social force model and the fundamentals of cellular automata and their applicability.

Multiagent Systems Lecture Introduction: Video

Link to the repositories: TBD

Previous microcredits: None

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K02 Intelligent Agents

Content:
The topic of intelligent agents includes models of intelligent autonomous agents, control, environments in multi-agent systems and an abstract agent architecture.

Learning Objectives:
Students are familiar with models and architectures of intelligent autonomous agents. They understand how agents deal with dynamic systems. They know different types of system environments/environments and can define them. They are able to formally describe the nature of agents as well as their tasks and system environment. They know the properties of reactive and state-based agents.

Multiagent Systems Lecture: Intelligent Agents Video

Link to the repositories: TBD

Previous microcredits: MAS-K01 Introduction to multiagent systems

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K03 Intelligent agents: Architectural approaches

Content:
Intelligent agents are based on different architectural approaches, such as: Deliberative architectures like Deductive Reasoning or STRIPS Planning, reactive architectures like the Subsumption architecture and Ant colony models as well as hybrid architectures.

Learning Objectives:
Students know the characteristics, advantages and disadvantages of the deliberative agent architecture and have experienced these in concrete examples using the STRIPS AI Planner. They are familiar with the STRIPS AI Planner and can formally implement their own scenarios. You know the properties of reactive agents. They have become familiar with the Subsumption Architecture. Students are familiar with the multi-agent architecture InteRRaP.

Multiagent Systems Lecture: Intelligent Agents: Architectural approaches Video

Link to the repositories: TBD

Previous microcredits: MAS-K02 Intelligent Agents

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K04 Practical Reasoning and BDI architectures

Content:
This part of the Multiagent Systems Module contains introductions to the topics of Intentionality, the Belief-Desire-Intention (BDI) architectures, practical reasoning and applicable algorithms. Two included concrete manifestations of BDI architectures are IRMA: Intelligent Resource-bounded Machine Architecture and PRS - Practical Reasoning System.

Learning Objectives:
Students know the basic features of practical reasoning. They understand the stylization of the Belief-Desire-Intention approach. They know implemented variants of practical reasoning agents.

Multiagent Systems Lecture: Practical Reasoning and BDI architectures: Video

Link to the repositories: TBD

Previous microcredits: MAS-K03 Intelligent agents: Architectural approaches

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K05 Modeling Concurrent Systems

Content:
In this chapter Agents as concurrent processes are being introduced, including motivation and basic definitions. Additionaly the Modeling of concurrent systems in the shape of Labeled transitions systems(LTS) and Finite State Processes(FSP) are taught. To conclude, concurrency and concurrent executions are presented, featuring deadlocks and the Dining Philosophers Problem (DPP).

Learning Objectives:
Students have knowledge of agent-oriented programming with concurrency. They know the properties of and differences between labeled transition systems (LTS) and finite state processes (FSP). They are aware of the technical challenges involved in implementing concurrent systems.

Multiagent Systems Lecture: Modeling Concurrent Systems: Video (Part 1), Video (Part 2)

Link to the repositories: TBD

Previous microcredits: MAS-K04 Practical Reasoning and BDI architectures

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K06 Multiagent Systems 1: Working Together

Content:
This training segment covers the coordination and cooperation in multi-agent systems. Cooperative distributed problem solving, the contract network protocol and simulated trading are elementary components.

Learning Objectives:
Students have visualized essential characteristics of coordination, cooperation and collaboration of automated systems. They have knowledge of Speech Act Theory and KQML. They have knowledge of Cooperative Distributed Problem Solving (CDPS) and the Contract Net Protocol (CNP).

Multiagent Systems Lecture: Working Together: Video (Part 1), Video (Part 2)

Link to the repositories: TBD

Previous microcredits: MAS-K05 Modeling Concurrent Systems

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K07 Interaction of individual rational agents: Basic concepts

Content:
This part will introduce the basic concepts of the interaction of individual rational agents. Further explained are normal form games (NFG), dominant strategies and the Nash equilibrium as well as pure and mixed strategies and the elimination of dominated strategies.

Learning Objectives:
Students have mastered the basics of game theory. They have experience with various game theory scenarios as well as pure and mixed strategies. They have the skills to identify and deal with dominant strategies.

Multiagent Systems Lecture: Interaction of individual rational agents: Basic concepts: Video

Link to the repositories: TBD

Previous microcredits: MAS-K06 Multiagent Systems 1: Working Together

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K08 Interaction of individual rational agents: Applications

Content:
This part continues the previous section (MAS-K07). The basic concepts of the interaction of individual rational agents are being complemented with exemplary applications and the concideration of the price of anarchy concept. Applications of game theory, N-player games and games with continuous payoffs add to this sections definition.

Learning Objectives:
Students have experience with the most important example scenarios in game theory. They are able to determine Nash equilibria.

Multiagent Systems Lecture: Interaction of individual rational agents: Applications: Video

Link to the repositories: TBD

Previous microcredits: MAS-K07 Interaction of individual rational agents: Basic concepts

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K09 Collective decision-making by automated agents

Content:
This lecture encloses the topic of Collective decision-making in multi-agent systems including the introduction to Choice Mechanisms, Social Choice and Social Welfare concepts.

Learning Objectives:
Students are familiar with computational social choice theory and its characteristics. They have knowledge of common choice mechanisms and their scope of competence. They have knowledge of parallels between game theory, computational social choice theory, decision theory and auctions.

Multiagent Systems Lecture: Collective decision-making by automated agents: Video

Link to the repositories: TBD

Previous microcredits: MAS-K08 Interaction of individual rational agents: Applications

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K10 Automated negotiation

Content:
The lecture Automated negotiations reviews ways of reaching agreements between agents by utilizing Negotiation mechanisms, suitable protocols and strategies. Additionally task-oriented problem domains are being introduced.

Learning Objectives:
Students have knowledge of negotiation models and their mechanisms. They are familiar with various negotiation protocols and can relate these to the appropriate problem domains.

Multiagent Systems Lecture: Automated negotiation: Video

Link to the repositories: TBD

Previous microcredits: MAS-K09 Collective decision-making by automated agents

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K11 Logic Programming

Content:
This lecture includes the Basics of logic programming. The foundation of the topic will be the Horn Logic which is going to be concluded by the Prolog programming language.

Learning Objectives:
Students are proficient in the ProLog logic programming language. They understand horn logic and are able to write programs based on predicate logic.

Multiagent Systems Lecture: Logic Programming: Video

Link to the repositories: TBD

Previous microcredits: MAS-K10 Automated negotiation

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal

MAS-K12 Agent-Oriented Programming

Content:
The last lecture of the Multiagent Systems module concludes with Agent-oriented programming. The focus lies on programming of BDI agents carried out with the instruments of AgentSpeak/Jason.

Learning Objectives:
Students are familiar with the AgentSpeak(L) programming language and are trained in using the Jason interpreter. They know the structure and components of the multi-agent systems developed with AgentSpeak(L). They are able to develop multi-agent systems independently.

Multiagent Systems Lecture: Agent-Oriented Programming: Video

Link to the repositories: TBD

Previous microcredits: MAS-K11 Logic Programming

Extent: 0.5 ECTS
Responsible: Institut für Informatik, TU Clausthal