Explaining BDI agent behaviour through dialogue

Abstract BDI agents act in response to external inputs and their internal plan library. Understanding the root cause of agent action is often difficult, this paper we present a dialogue based approach for explaining behaviour agent. We consider two participants who may have different views regarding beliefs, plans events which drove (encoded via traces). These make utterances incrementally reveal traces each other, allowing them identify divergences traces, or conclude that agree. In practice, envision human taking on role participant, with itself acting as other participant. The then facilitates explanation, understanding debugging behaviour. After presenting our formalism its properties, describe implementation system provide an example use simple scenario.