Value Compression of Pattern Databases
Abstract
One common pattern database compression technique is to merge adjacent database entries and store the minimum of merged entries to maintain heuristic admissibility. In this paper we propose a compression technique that preserves every entry, but reduces the number of bits used to store each entry, therefore limiting the values that can be represented. Even when this technique throws away low values in the heuristic, it can still have better performance than the traditional approach. We develop a theoretical basis for selecting which values to keep and show improved performance in both unidirectional and bidirectional search.
Cite
Text
Sturtevant et al. "Value Compression of Pattern Databases." AAAI Conference on Artificial Intelligence, 2017. doi:10.1609/AAAI.V31I1.10665Markdown
[Sturtevant et al. "Value Compression of Pattern Databases." AAAI Conference on Artificial Intelligence, 2017.](https://mlanthology.org/aaai/2017/sturtevant2017aaai-value/) doi:10.1609/AAAI.V31I1.10665BibTeX
@inproceedings{sturtevant2017aaai-value,
title = {{Value Compression of Pattern Databases}},
author = {Sturtevant, Nathan R. and Felner, Ariel and Helmert, Malte},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2017},
pages = {912-918},
doi = {10.1609/AAAI.V31I1.10665},
url = {https://mlanthology.org/aaai/2017/sturtevant2017aaai-value/}
}