Solution-based phase saving for CP: a value-selection heuristic to simulate local search behavior in complete solvers

Emir Demirović, Geoffrey Chu, Peter J. Stuckey

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

10 Citations (Scopus)


Large neighbourhood search, a meta-heuristic, has proven to be successful on a wide range of optimisation problems. The algorithm repeatedly generates and searches through a neighbourhood around the current best solution. Thus, it finds increasingly better solutions by solving a series of simplified problems, all of which are related to the current best solution. In this paper, we show that significant benefits can be obtained by simulating local-search behaviour in constraint programming by using phase saving based on the best solution found so far during the search, activity-based search (VSIDS), and nogood learning. The approach is highly effective despite its simplicity, improving the highest scoring solver, Chuffed, in the free category of the MiniZinc Challenge 2017, and can be easily integrated into modern constraint programming solvers. We validated the results on a wide range of benchmarks from the competition library, comparing against seventeen state-of-the-art solvers.

Original languageEnglish
Title of host publicationPrinciples and Practice of Constraint Programming
Subtitle of host publication24th International Conference, CP 2018 Lille, France, August 27–31, 2018 Proceedings
EditorsJohn Hooker
Place of PublicationCham Switzerland
Number of pages10
ISBN (Electronic)9783319983349
ISBN (Print)9783319983332
Publication statusPublished - 2018
Externally publishedYes
EventInternational Conference on Principles and Practice of Constraint Programming 2018 - Lille, France
Duration: 27 Aug 201831 Aug 2018
Conference number: 24th

Publication series

NameLecture Notes in Computer Science
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


ConferenceInternational Conference on Principles and Practice of Constraint Programming 2018
Abbreviated titleCP 2018
Internet address

Cite this