Argument-based Cooperative Decision-Making for Autonomous Vehicles

Abstract

In this paper, we focus on multi-agent multi-criteria decision problems in dynamic environments where autonomous vehicles need to cooperate to make decisions. When a group of agents faces these problems, each agent may have a different preference about which criteria are more relevant. Hence, agents’ preference rules will be used to generate arguments that represent all the agents’ points of view. Our approach performs an argumentation-based analysis to entail the best alternatives, and explains how it arrives at its conclusions. We provide various scenarios to show how our proposal behaves in different situations. We formally show some results about how our current proposal behaves in dynamic environments regarding the computation of the solution of the problem.