SibylSat: Using SAT as an Oracle to Perform a Greedy Search on TOHTN Planning
This work addresses a specific bottleneck in HTN planning for AI researchers, offering an incremental improvement over existing SAT-based methods.
The paper tackles the problem of efficiently solving totally-ordered HTN planning problems by introducing SibylSat, a novel SAT-based method that uses a greedy search approach with a heuristic derived from solving a relaxed SAT problem. The result shows that SibylSat outperforms existing SAT-based TOHTN approaches in runtime and plan quality on most IPC benchmarks, solving a larger number of problems.
This paper presents SibylSat, a novel SAT-based method designed to efficiently solve totally-ordered HTN problems (TOHTN). In contrast to prevailing SAT-based HTN planners that employ a breadth-first search strategy, SibylSat adopts a greedy search approach, enabling it to identify promising decompositions for expansion. The selection process is facilitated by a heuristic derived from solving a relaxed problem, which is also expressed as a SAT problem. Our experimental evaluations demonstrate that SibylSat outperforms existing SAT-based TOHTN approaches in terms of both runtime and plan quality on most of the IPC benchmarks, while also solving a larger number of problems.