Decision Making with Assumption-Based Argumentation

Abstract

In this paper, we present two different formal frameworks for representing decision making. In both frameworks, decisions have multiple attributes and meet different goals. In the second framework, decisions take into account preferences over goals. We also study a family of decision functions representing making decisions with different criteria, including decisions meeting all goals, most goals, goals no other decisions meet, and most preferred achievable goals. For each decision function, we define an argumentation-based computational mechanism for computing and explaining the selected decisions. We make connections between decision making and argumentation semantics, i.e., selected decisions in a decision making framework are admissible arguments in the corresponding argumentation framework. The main advantage of our approach is that it not only selects decisions but also gives an argumentation-based justification of selected decisions.