Encoding Partial Plans for Heuristic Search

Encoding Partial Plans for Heuristic Search. Bercher, P. & Biundo, S. In Proceedings of the 4th Workshop on Knowledge Engineering for Planning and Scheduling (KEPS 2013), pages 11–15, 2013.

Paper

Slides abstract bibtex 3 downloads

We propose a technique that allows any planning system that searches in the space of partial plans to make use of heuristics from the literature which are based on search in the space of states. The technique uses a problem encoding that reduces the problem of finding a heuristic value for a partial plan to finding a heuristic value for a state: It encodes a partial plan into a new planning problem, s.t. solutions for the new problem correspond to solutions reachable from the partial plan. Evaluating the goal distance of the partial plan then corresponds to evaluating the goal distance of the initial state in the new planning problem.

@InProceedings{Bercher13EncodingPlans,
  author    = {Pascal Bercher and Susanne Biundo},
  title     = {Encoding Partial Plans for Heuristic Search},
  booktitle = {Proceedings of the 4th Workshop on Knowledge Engineering for Planning and Scheduling ({KEPS} 2013)},
  year      = {2013},
  pages     = {11--15},
  abstract  = {We propose a technique that allows any planning system that searches in the space of partial plans to make use of heuristics from the literature which are based on search in the space of states. The technique uses a problem encoding that reduces the problem of finding a heuristic value for a partial plan to finding a heuristic value for a state: It encodes a partial plan into a new planning problem, s.t. solutions for the new problem correspond to solutions reachable from the partial plan. Evaluating the goal distance of the partial plan then corresponds to evaluating the goal distance of the initial state in the new planning problem.},
  url_Paper  = {https://bercher.net/publications/2013/Bercher2013EncodingPlans.pdf},
  url_Slides = {https://bercher.net/publications/2013/Bercher2013EncodingPlansSlides.pdf},
  keywords   = {workshop}
}

Downloads: 3