Parallel search-based methods in optimization
Search-based methods like Branch and Bound and Branch and Cut are essential tools in solving difficult problems to optimality in the field of combinatorial optimization, and much experience has been gathered regarding the design and implementation of parallel methods in this field.
Search-based methods appear, however, also in connection with certain continuous optimization problems and problems in Artificial Intelligence, and parallel versions hereof have also been studied.
Based on parallel Branch and Bound algorithms, the advantages as well as the difficulties and pitfalls in connection with parallel search-based methods are outlined. Experiences with parallel search-based methods from the three fields mentioned are then described and compared in order to reveal similarities as well as differences across the fields.
Unable to display preview. Download preview PDF.
- 1.S. Berner, “Ein paralleles Verfahren zur verifizierten globalen Optimierung”, Fachbereich Mathematik, Univ. Wuppertal, 1995.Google Scholar
- 2.O. Caprani, B. Godthaab and K. Madsen, “Use of a real-values local minimum in parallel interval global optimization”, Interval Computations 2, p. 71–82, 1993.Google Scholar
- 3.J. Clausen, “Parallel Branch-and-Bound — Principles and Personal experiences”, 20 p., presented at Nordic Summer Course on Parallel Algorithms in Mathematical Programming, Linköping, August 1995. DIKU report 95/29, to be published by Kluwer.Google Scholar
- 4.J. Clausen and J.L. Träff, “Implementation of Parallel Branch-and-Bound Algorithms — Experiences with the Graph Partitioning Problem”, Annals of Operations Research 33 p. 331–349, 1991.Google Scholar
- 5.J. Clausen and J.L. Träff, “Do Inherently Sequential Branch-and-Bound Algorithms Exist ?”, Parallel Processing Letters 4, 1–2, p. 3–13, 1994.Google Scholar
- 6.J. Clausen and M. Perregaard, “Solving Large Quadratic Assignment Problems in Parallel”, DIKU report 1994/22, 14 p., to appear in Computational Optimization and Applications.Google Scholar
- 7.J. Clausen and Michael Perregaard, “On the Best Search Strategy in Parallel Branch-and-Bound — Best-First-Search vs. Lazy Depth-First-Search”, DIKU report 96/14, 11 pages.Google Scholar
- 8.L.C. Dixon and M. Jha, “Parallel algorithms for global optimization”, Journal of Optimization Theory and Applications 79, p. 385–395, 1993.Google Scholar
- 9.J. Eriksson, “Parallel Global Optimization Using Interval Analysis”, Dissertation, University of Umeå, Sweden, 1991.Google Scholar
- 10.T. Henriksen and K. Madsen, “Use of a depth-first strategy in parallel global optimization”, Report no. 92-10, Inst. for Numerical Analysis, Tech. Univ. of Denmark, 1992.Google Scholar
- 11.B. Gendron and T. G. Cranic, “Parallel Branch-and-Bound Algorithms: Survey and Synthesis”, Operations Research 42 (6), p. 1042–1066, 1994.Google Scholar
- 12.A. Grama and V. Kumar, “Parallel Search algorithms for Discrete Optimization Problems”, ORSA Journal of Computing 7 (4), p. 365–385, 1995.Google Scholar
- 13.R.E. Hansen, E. Hansen and A. Leclerc, “Rigorous methods for global optimization”, in C.A. Floudas and P.M. Pardalos (eds.) “Recent Advances in Global Optimization”, Princeton University Press, 1992.Google Scholar
- 14.R. Lüling, B. Monien, A. Reinefeld and S. Tshöke, “Mapping Tree-Structured Combinatorial Optimization Problems onto Parallel Computers” in A. Ferreira and P.M. Pardalos (eds.) “Solving Combinatorial Optimization Problems in Parallel — Methods and Techniques”, Springer LNCS 1054, 1996.Google Scholar
- 15.M. Perregaard and J. Clausen, “Solving Large Job Shop Scheduling Problems in Parallel”, DIKU report 94/35, under revision for Annals of OR.Google Scholar