search

nathansttt / hog2

Pathfinding and search testbed/visualization suite. Current code is in PDB-refactor branch.
MIT License
107 stars 54 forks source link

What is the function of "bucket" in the scenario? #36

Closed pineappleKID closed 3 years ago

pineappleKID commented 3 years ago

As introduced in https://movingai.com/benchmarks/formats.html, the first column of .scen file means "bucket". But there is no explanation about "bucket". "bucket" is used in ScenarioLoader.cpp and ScenarioLoader.h but I still have no idea how to use "bucket". https://github.com/nathansttt/hog2/blob/PDB-refactor/apps/canonicalGrids/Sample.cpp will do something if bucket > 0 but the value of bucket is not used.

I notice that, in general, the greater bucket is, the greater optimal path length will be but this is not always true in the scenario file.

Thanks in advance.

nathansttt commented 3 years ago

Buckets are just the optimal path length divided by 4. It's used for grouping similar-length paths together for statistical purposes - to get a larger sample size. My lab has been using this approach since this paper: https://webdocs.cs.ualberta.ca/~nathanst/papers/partialpathfinding.pdf