Experimental evaluation of sBELA0/sBELA* (IJCAI 2026)
收藏Zenodo2026-05-14 更新2026-05-29 收录
下载链接:
https://zenodo.org/doi/10.5281/zenodo.20184563
下载链接
链接失效反馈官方服务:
资源简介:
Data files with the information required to replicate all the experiments reported in the paper:
Linares López, Carlos; Herman, Ian, 2026. Finding Simple Shortest-Paths via Centroids. Proceedings of the 35th International Joint Conference on Artificial Intelligence (IJCAI 2026). Bremen (Germany).
This record contains the test and data files needed to test any of the algorithms implemented in the different domains (maps, roadmap and n-pancake). While these files can be used to run the experiments manually, there are different scripts available in the github repository "ksearch" which automate the experimentation, recollection of data and presentation of results under the directory "scripts".
Files are structured as follows:
random512: data+test files used in the maps domain:
random512-*-0.map contains the map whose contents are explained in the movingai repository.
random512-*.test contains the testcases for for the maps domain.
random512 .co/.gr files: represent the same maps but they are formatted according to the rules of the 9th DIMACS Implementation Challenge Shortest Path. These are necessary only for performing comparisons with other solvers that only process that format.
random512*unit.euclidean.test: test cases for the maps compiled into .gr/.co files.
test-{10, 20}.test: data files used in the {10, 20}-Pancake
USA-road.d.*.test: test files used in the roadmap domain.
To be able to run experiments in the roadmap domain it is necessary to download the .gr and .co files from 9th DIMACS Implementation Challenge: Shortest Paths
The format of all the data files is explained in the README.md files of the "ksearch" github repo under the corresponding directory.
All tests were randomly generated in previous research. The criteria for the random selection is explained in the paper:
Linares López, Carlos; Herman, Ian, 2024. Evolving A* to Efficiently Solve the k Shortest-Path Problem. Proceedings of the Twenty-Seventh European Conference on Artificial Intelligence (ECAI 2024), pages 4352-4359. Santiago de Compostela (Spain).
提供机构:
Zenodo
创建时间:
2026-05-14



