The search functionality is under construction.

IEICE TRANSACTIONS on Information

Solving Open Job-Shop Scheduling Problems by SAT Encoding

Miyuki KOSHIMURA, Hidetomo NABESHIMA, Hiroshi FUJITA, Ryuzo HASEGAWA

  • Full Text Views

    0

  • Cite this

Summary :

This paper tries to solve open Job-Shop Scheduling Problems (JSSP) by translating them into Boolean Satisfiability Testing Problems (SAT). The encoding method is essentially the same as the one proposed by Crawford and Baker. The open problems are ABZ8, ABZ9, YN1, YN2, YN3, and YN4. We proved that the best known upper bounds 678 of ABZ9 and 884 of YN1 are indeed optimal. We also improved the upper bound of YN2 and lower bounds of ABZ8, YN2, YN3 and YN4.

Publication
IEICE TRANSACTIONS on Information Vol.E93-D No.8 pp.2316-2318
Publication Date
2010/08/01
Publicized
Online ISSN
1745-1361
DOI
10.1587/transinf.E93.D.2316
Type of Manuscript
LETTER
Category
Artificial Intelligence, Data Mining

Authors

Keyword