Modellering och Syntes av kunskapsbaserade strategier för asynkrona multiagentsystem

This thesis explores strategy synthesis in the nascent Asynchronous Multi-Agent System (AMAS), which is a system where multiple players acting asynchronously and with imperfect information collaborate in order to achieve a goal. Previous work has proven a heuristic for knowledge-based strategy synthesis in a related system where the players act synchronously. We propose an adaptation of said heuristic to AMAS. The adapted heuristic is then tested by applying it to two modified models of an example AMAS from a previous paper, one designed to test the heuristic’s capability of finding a surely winning knowledge-based strategy, and one to test it when no such strategy can exist. We find that the proposed heuristic behaves as expected in both cases.