Argumentation framework based on evidence theory

In many fields of automated information processing it becomes crucial to consider imprecise, uncertain or inconsistent pieces of information. Therefore, integrating uncertainty factors in argumentation theory is of paramount importance. Recently, several argumentation based approaches have emerged to model uncertain data with probabilities. In this paper, we propose a new argumentation system called evidential argumentation framework that takes into account imprecision and uncertainty modeled by means of evidence theory. Indeed, evidence theory brings new semantics since arguments represent expert opinions with several weighted alternatives. Then, the evidential argumentation framework is studied in the light of both Smets and Demspter-Shafer interpretations of evidence theory. For each interpretation, we generalize Dung’s standard semantics with illustrative examples. We also investigate several preference criteria for pairwise comparison of extensions in order to select the ones that represent potential solutions to a given decision making problem.