Applying Local Search to Disjunctive Temporal Problems
Abstract
We present a method for applying local search to overconstrained instances of the Disjunctive Temporal Problem (DTP). Our objective is to generate high quality solutions (i.e., solutions that violate few constraints) in as little time as possible. The technique presented here differs markedly from previous work on DTPs, as it operates within the total assignment space of the underlying CSP rather than the partial assignment space of the related meta-CSP. We provide experimental results demonstrating that the use of local search leads to substantially improved performance over systematic methods.
Cite
Text
Moffitt and Pollack. "Applying Local Search to Disjunctive Temporal Problems." International Joint Conference on Artificial Intelligence, 2005.Markdown
[Moffitt and Pollack. "Applying Local Search to Disjunctive Temporal Problems." International Joint Conference on Artificial Intelligence, 2005.](https://mlanthology.org/ijcai/2005/moffitt2005ijcai-applying/)BibTeX
@inproceedings{moffitt2005ijcai-applying,
title = {{Applying Local Search to Disjunctive Temporal Problems}},
author = {Moffitt, Michael D. and Pollack, Martha E.},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2005},
pages = {242-247},
url = {https://mlanthology.org/ijcai/2005/moffitt2005ijcai-applying/}
}