Prioritizing Answer Sets Based on Conditional Expert Knowledge

Abstract

Answer set programming (ASP) and conditional reasoning both are powerful and widely used methodologies from the field of knowledge representation and reasoning (KR) which are capable of formalizing default statements that usually hold but also leave room for exceptions. While ASP convinces with an intuitive rule-based syntax and fast solvers, conditionals come along with a sophisticated preference-based semantics. Here, we combine both approaches by calculating answer sets which we then prioritize based on conditional expert knowledge. We apply our hybrid approach to the task of planning warehouse layouts from the logistics domain which is predestinated for our approach because it involves, on the low-level, many variables and technical framework conditions (like rack positions) and, on the high-level, expert knowledge of the layout designer.