Lifted Heuristics for Timeline-Based Planning
De Benedictis Riccardo, Amedeo Cesta · Frontiers in artificial intelligence and applications · 2020
This paper discusses the issue of efficient resolution of timeline-based planning problems. In particular, taking inspiration from the more classical heuristics for the resolution of STRIPS-like problems, it proposes a new heuristic strategy which, while maintaining the variables lifted, allows more accurate decisions. The concepts presented in this work pave the way for a new type of heuristics which, at present, allow this kind of solvers a significant performance improvement.