Explaining Plan Selection Through an Analysis of Object Transition Sequences

Creating explanations that justify why a particular plan has been selected is an important goal of explainable planning. However, as each selected plan sequence will implicitly encapsulate the balance that has been made between dependency, constraint and choice, the job of justifying a plan is challenging. A powerful concept in explainable planning is to use contrastive explanations; however, this requires some method of accessing suitable alternatives. In this work we focus on the transition sequences of specific objects in a plan and define a set of alternatives that may have been used instead. We then analyse the choice that was made: we attempt to generate an explanation to justify the used transition sequence and explore the viability of using alternative options within a plan. We discuss our observations from preliminary component testing.