Abstract
In everyday life human decision-making is often based on arguments and counter-arguments. Decisions made in this way have a basis that can be easily referred to for explanation purposes as not only is a best choice suggested, but also the reasons of this recommendation can be provided in a format that is easy to grasp. The formal foundations of argumentation have been well explored in the academic literature, but in contrast the actual implemented computational models have been less widely investigated. Research into the practicality of argumentation has also lacked the development of empirical evaluation techniques, which due to the well-documented computational complexity issues associated with argumentative reasoning, could prove a major stumbling block in gaining mainstream acceptance. Our core motivation of this thesis is contributing to raising awareness that research on models of argumentation is on the cusp of a move into a phase of engineering applicable implementations. Accordingly, we present "Argue tuProlog", an open source argumentation-based reasoning engine, which is rigorous in design and implementation, and also "ArgueGen", a knowledgebase generator which forms part of an empirical evaluation platform and methodology for experimenting and benchmarking our engine implementation. One of the key goals of the work presented in this thesis is answering the questions "Is conducting empirical evaluation of an argumentation-based reasoning engine advantageous to the design and implementation of the engine?" and "What benefits are provided by performing empirical evaluation of an engine?" Our experiments with AtuP2 and ArgueGen demonstrate that the answer to the first question is a resounding yes, and the benefits offered by such an evaluation include not only the indication of specific areas of the engine implementation to target in order to improve performance or reduce the inherent computation expense of arguing, but also the tangible evidence as to their success or failure.