Limitations of Front-to-End Bidirectional Heuristic Search

Abstract

We present an intuitive explanation for the limited effectiveness of front-to-end bidirectional heuristic search, supported with extensive evidence from many commonly-studied domains. While previous work has proved the limitations of specific algorithms, we show that any front-to-end bidirectional heuristic search algorithm will likely be dominated by unidirectional heuristic search or bidirectional brute-force search. We also demonstrate a pathological case where bidirectional heuristic search is the dominant algorithm, so a stronger claim cannot be made. Finally, we show that on the four-peg Towers Of Hanoi with arbitrary start and goal states, bidirectional brute-force search outperforms unidirectional heuristic search using pattern-database heuristics.

Cite

Text

Barker and Korf. "Limitations of Front-to-End Bidirectional Heuristic Search." AAAI Conference on Artificial Intelligence, 2015. doi:10.1609/AAAI.V29I1.9374

Markdown

[Barker and Korf. "Limitations of Front-to-End Bidirectional Heuristic Search." AAAI Conference on Artificial Intelligence, 2015.](https://mlanthology.org/aaai/2015/barker2015aaai-limitations/) doi:10.1609/AAAI.V29I1.9374

BibTeX

@inproceedings{barker2015aaai-limitations,
  title     = {{Limitations of Front-to-End Bidirectional Heuristic Search}},
  author    = {Barker, Joseph Kelly and Korf, Richard E.},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2015},
  pages     = {1086-1092},
  doi       = {10.1609/AAAI.V29I1.9374},
  url       = {https://mlanthology.org/aaai/2015/barker2015aaai-limitations/}
}