DMCA
Specializing Russian Doll Search (2001)
| Venue: | In Proceedings of CP |
| Citations: | 12 - 0 self |
BibTeX
@INPROCEEDINGS{Meseguer01specializingrussian,
author = {Pedro Meseguer and Martì Sánchez},
title = {Specializing Russian Doll Search},
booktitle = {In Proceedings of CP},
year = {2001},
pages = {464--478},
publisher = {Springer Verlag}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Russian Doll Search (RDS) is a clever procedure to solve overconstrained problems. RDS solves a sequence of nested subproblems, where two consecutive subproblems differ in one variable only. We present the Specialized RDS (SRDS) algorithm, which solves the current subproblem for each value of the new variable with respect to the previous subproblem. The SRDS lower bound is superior to the RDS lower bound, which allows for a higher level of value pruning, although more work per node is required. Experimental results on random and real problems show that this extra work is often beneficial, providing substantial savings in the global computational effort.
