RT Journal Article
T1 Enhancing set constraint solvers with bound consistency
A1 Correas Fernández, Jesús
A1 Estévez Martín, Sonia
A1 Sáenz Pérez, Fernando
A2 Lin, Binshan
AB Constraint solvers have proved to be a powerful tool for both expressing and solving complex problems in many contexts, in particular in Automated Planning and Scheduling. In this work, we propose new constraints to enhancing an implementation of a constraint solver for sets of integers (ic_sets) embedded in the ECLiPSe system. This way, we confer a new vision of modelling that in particular allows a task to be planned on alternate days. In addition, these new constraints make finite domain and finite sets domains to cooperate, reducing the search space and improving the efficiency of the existing set solvers for some specific cases. We report on a prototype implementation and apply these constraints to modelling scheduling problems as benchmarks for a performance analysis. Experiments reveal a relevant improvement with respect to the standard ic_sets solver which only features cardinality as cooperation mechanism between domains. Both enhanced expressiveness and performance are key points for practical applications of such solvers.
PB Elsevier
SN 0957-4174
YR 2018
FD 2018-02-01
LK https://hdl.handle.net/20.500.14352/93814
UL https://hdl.handle.net/20.500.14352/93814
LA eng
NO J. Correas Fenández, S. Estévez-Martín, and F. Sáenz-Pérez. Enhancing set constraint solvers with bound consistency. Expert Systems with Applications, 92:485 - 494, 2018.
DS Docta Complutense
RD 22 abr 2025