Многоагентные системы (Multi-Agent Systems)

In an increasingly networked environment with increasingly autonomous agents, the study of decentralized algorithms for coordination of collaborating or competing agents is of great importance. The MAS course is dedicated to the study of a wide range of issues surrounding multi-agent interactions. We will use games as a highly suitable formal model to study such interactions, with examples ranging from abstract games to real-world applications. The module will teach both the mathematical basics as well as practical algorithms in real-world applications. Topics will include game theory (equilibria, repeated games, bayesian games), game tree search, communication, negotiation, argumentation, mechanism design, multi-agent learning.