Pushing the Envelope in Numeric Pattern Planning

In this paper, we present a symbolic search-based procedure for numeric planning based on Symbolic Pattern Planning (SPP). In SPP, a pattern is a sequence of actions used to define a logic formula whose models correspond to sequences of applicable actions and reachable states. Here, starting from the empty pattern, we iteratively extend and compress it using search techniques until a goal state is reached. We prove the correctness and completeness of the procedure and demonstrate its good performance compared to both the original SPP approach and other publicly available numeric planners on the 2023 International Planning Competition Agile track.