Multi-Agent Decision Making with Assumption-based Argumentation

Much research has been devoted in recent years to argumentationbased decision making. However, less attention has been given to argumentation-based decision making amongst multiple agents. We present a multi-agent decision framework based on Assumptionbased Argumentation. In our model, agents have goals and decisions have attributes which satisfy goals. Our framework supports agents with different goals, candidate decisions, attributes and relations amongst them. Using an existing argumentation-based dialogue framework, we show how two agents can argue towards “good” decisions in a distributed manner. We show that, under specific conditions, “good” decisions correspond to (1) admissible arguments for the two agents and (2) claims of successful dialogues between the two agents. Thus, this work connects decision making with multi-agent argumentation and dialogues.