We define the weighted-sequence problem inspired by an important industrial problem in oracle query optimization. We evaluate several approaches to solving this problem using answer set programming (ASP) system clingo and constraint answer set programming (CASP) solver clingcon. The focus of this paper is an experimental evaluation of search procedures behind clingo and clingcon. We used instances of the weighted-sequence problem where integer parameters were quite small so that grounding was not an issue and we could focus on comparison of the effectiveness of the clingo and clingcon solvers. Our key finding is that the effectiveness of the search procedure used by clingcon lags behind that of clingo. It points to the need for further research on the integration of ASP and CSP technologies.
|State||Published - 2011|
|Event||4th Workshop on Answer Set Programming and Other Computing Paradigms, ASPOCP 2011, collocated with the 27th International Conference on Logic Programming, ICLP 2011 - Lexington, United States|
Duration: Jul 10 2011 → …
|Conference||4th Workshop on Answer Set Programming and Other Computing Paradigms, ASPOCP 2011, collocated with the 27th International Conference on Logic Programming, ICLP 2011|
|Period||7/10/11 → …|
Bibliographical noteFunding Information:
We are grateful to Philip Cannata for bringing the problem of finding an optimal join order in query optimization to our attention, and to Vladimir Lifschitz and Yuanlin Zhang for useful discussions related to the topic of this work. Yuliya Lierler was supported by a CRA/NSF 2010 Computing Innovation Fellowship, Miroslaw Truszczynski by the NSF grant IIS-0913459, and Shaden Smith and Alex Westlund by the NSF REU Supplement to that grant.
© 2011 Computing Research Repository (CoRR). All rights reserved.
ASJC Scopus subject areas
- Theoretical Computer Science
- Computational Theory and Mathematics
- Hardware and Architecture
- Artificial Intelligence