Can We Change the Game? Reasoning about Dynamic Multi-Agent Systems
Most research on logics for strategic reasoning in Multi-Agent Systems (MAS) has traditionally focused on static models, such as concurrent game structures, which represent a fixed set of system configurations and the transitions between them. Such models are unable to capture scenarios where the system’s structure undergoes dynamic changes, whether triggered by agent actions or caused by an external factor. As MAS increasingly operate in dynamic and unpredictable settings, their design and maintenance require reasoning about how structural changes affect system behavior. When existing models fail to produce desirable outcomes or are incorrect, computing repairs offers an alternative to complete redesigns. This talk explores recent advances in logic-based methods for dynamic MAS, including reasoning about model modifications and repairing flawed games.