The problem was solved using the branch-and-bound algorithm described in the paper "Solving quadratic assignment problems using convex quadratic programming relaxations," N.W. Brixius and K.M. Anstreicher, available at http://www.biz.uiowa.edu/faculty/anstreicher/qapqp2.ps. The algorithm was implemented using the Master-Worker (MW) distributed processing system (see http://www.mcs.anl.gov/metaneos/papers/mw2.ps), developed as a joint collaboration between researchers at the University of Wisconsin and Argonne National Labs as part of the MetaNEOS project (see http://www.mcs.anl.gov/metaneos/).
The computation was performed on a pool of workstations using the Condor high-throughput computing system developed at the University of Wisconsin (see http://www.cs.wisc.edu/condor/) in a total wall time of approximately 4 days, 8 hours. During this time the number of active worker machines averaged approximately 200. Machines from the University of Wisconsin, the Albuquerque High Performance Computing Center, and the Italian Istituto Nazionale di Fisica Nucleare (INFN) all participated in the computation. A total of 6.58E7 CPU seconds was expended by worker machines. The equivalent computation time on a single HP-C3000 workstation would be approximately 435 days.
A permutation with an objective value of 5166 was found by the simulated annealing code of E. Taillard (available from http://sunst50.einev.ch/prive/pro/taillard/). The branch-and-bound computation was initialized with an incumbent objective value of 5167, and required 2,935,394,013 nodes. The B&B process verified that 5166 is the optimal value for the problem, and found the following permutation with this value:
18 21 9 1 28 20 11 3 13 12 10 19 14 22 15 2 25 16 4 23 7 17 24 26 5 27 8 6
Full details of the distributed branch-and-bound implementation using the MW system will be given in a paper under preparation.
Kurt Anstreicher, University of Iowa Nathan Brixius, University of Iowa Jean-Pierre Goux, Northwestern University and Argonne National Labs Jeff Linderoth, Argonne National Labs
From: Kurt AnstreicherDear colleagues: We are pleased to announce the exact solution of the "nug30" quadratic assignment problem (QAP). This problem was first posed in the paper C.E. Nugent, T.E. Vollman, and J. Ruml, "An experimental comparison of techniques for the assignment of facilities to locations," Operations Research 16 (1968) pp. 150-173, and has long been considered an outstanding open "challenge problem" in the optimization literature. The problem data is available from QAPLIB (http://www.imm.dtu.dk/~sk/qaplib). The problem was solved using the branch-and-bound algorithm described in the paper "Solving quadratic assignment problems using convex quadratic programming relaxations," N.W. Brixius and K.M. Anstreicher, available at http://www.biz.uiowa.edu/faculty/anstreicher/qapqp2.ps. It was implemented using the Master-Worker (MW) distributed processing system developed in the metaNEOS project and described in the paper "An enabling framework for master-worker computing applications on the computational grid," J.-P. Goux, S. Kulkarni, J. Linderoth, and M. Yoder, available from http://www.mcs.anl.gov/metaneos/papers/mw2.ps. The computation was performed on a dynamic grid of workstations, massively parallel processors, and other CPUs using the Condor high-throughput computing system developed at the University of Wisconsin, together with tools from the Globus toolkit. The total wall-clock time was approximately 6.9 days. The number of worker machines averaged approximately 650, and peaked at 1009. Machines from the following institutions participated in the computation (listed in order of CPU seconds contributed): University of Wisconsin, Argonne National Laboratory, Georgia Institute of Technology Interactive High Performance Computing Laboratory, National Center for Supercomputing Applications (NCSA), Italian Istituto Nazionale di Fisica Nucleare (INFN), Albuquerque High Performance Computing Center, Northwestern University, Columbia University. A total of 3.46E8 CPU seconds was expended by worker machines. The equivalent computation time on a single HP-C3000 workstation would be approximately 6.9 years. According to QAPLIB a permutation with an objective value of 6124 was first obtained by J. Skorin-Kapov, see J. Skorin-Kapov, "Tabu search applied to the quadratic assignment problem," ORSA J. Computing 2 (1990), pp. 33-45. The branch-and-bound computation was initialized with an incumbent objective value of 6126, and required 11,892,208,412 nodes. The B&B process verified that 6124 is the optimal value for the problem, and found the following permutation with this value: 14,5,28,24,1,3,16,15,10,9,21,2,4,29,25,22,13,26,17,30,6,20,19,8,18,7,27,12,1 1,23 Full details of the distributed branch-and-bound implementation using the MW system will be given in a paper under preparation. Further information can be found at the following URLs: Condor: http://www.cs.wisc.edu/condor/ metaNEOS: http://www.mcs.anl.gov/metaneos/ MW : http://www.cs.wisc.edu/condor/mw Globus: http://www.globus.org/ We are extremely grateful to Steve Wright of Argonne National Lab, and Miron Livny of the University of Wisconsin, for their ongoing support of this research. Kurt Anstreicher, University of Iowa Nathan Brixius, University of Iowa Jean-Pierre Goux, Northwestern University and Argonne National Lab Jeff Linderoth, Argonne National Lab
Dear Colleagues, The Krarup 30b instance of the Quadratic Assignment Problem was just solved at the University of Pennsylvania using the dual procedure branch-and-bound algorithm (Hahn-Grant bound) in the equivalent of 182 days (15,737,136 seconds) on a single cpu HP-3000 workstation. Actual runtimes were; 13,806,183 seconds on the HP-3000 3,861,905 seconds on a 350 MHz UltraSparc 10 (about 1/2 as fast as the HP-3000) A total of 183,659,980 nodes were evaluated. This compares to the parallel solution of this problem instance by Kurt Anstreicher, University of Iowa Nathan Brixius, University of Iowa Jean-Pierre Goux, Northwestern University and Argonne National Lab Jeff Linderoth, Argonne National Lab which took the equivalent of 2.7 years on a single cpu HP-3000 workstation and required the evaluation of 5,136,036,412 nodes. Peter Hahn Systems Engineering Department School of Engineering and Applied Science Monique Guignard-Spielberg Operations and Information Management Department The Wharton School
Announcing the technical report: Maximum Stable Set Formulations and Heuristics Based on Continuous Optimization By Sam Burer, Renato Monteiro, Yin Zhang Technical Report TR00-34 Department of Computational and Applied Mathematics, Rice University, Houston, Texas 77005 The postscript file can be obtained from: http://www.caam.rice.edu/~zhang/reports/tr0034.ps Thank you. -- Yin Zhang zhang@caam.rice.edu Abstract: The stability number for a given graph $G$ is the size of a maximum stable set in $G$. The Lovasz theta number provides an upper bound on the stability number and can be computed as the optimal value of the Lovasz semidefinite program. In this paper, we show that restricting the matrix variable in the Lovasz semidefinite program to be rank-one or rank-two yields a pair of continuous, nonlinear optimization problems each having the global optimal value equal to the stability number. We propose heuristics for obtaining large stable sets in $G$ based on these new formulations and present computational results indicating the effectiveness of the heuristics.
Announcing a new report.... Computational experience with Stable Set Relaxations Gerald Gruber, Franz Rendl Research Report, Department of Mathematics, University of Klagenfurt, Austria, December 2000 We investigate relaxations for the maximum stable set problem based on the Lov$\acute{{\rm a}}$sz number $\vartheta (G)$ as an initial upper bound. We strengthen this relaxation by adding two classes of cutting planes, odd circuit and triangle inequalities. We present computational results using this tighter model on many classes of graphs. URL of the dvi file: - URL of the ps file: http://www.uni-klu.ac.at/groups/math/optimization/pub_en.html/ Contact: gerald.gruber@uni-klu.ac.at
From Tamas Terlaky, terlaky@mcmaster.ca Sat Dec 23 11:09:15 2000 Date: Wed, 29 Nov 2000 23:42:06 -0500 Subject: Conference announcement 1st Annual McMaster Optimization Conference: Theory and Applications (MOPTA 01) August 2-4, 2001, McMaster University Hamilton, Ontario, Canada The 1st annual McMaster Optimization Conference, (MOPTA 01) will be held at the campus of McMaster University. It will be hosted by the Advanced Optimization Lab at the Department of Computing and Software. SCOPE The conference is planned as an annual event aiming to bring together a diverse group of people from both discrete and continuous optimization, working on both theoretical and applied aspects. The format will consist of a small number of invited talks from distinguished speakers and a set of selected contributed talks, spread over three days with no parallel sessions. Our target is to present a diverse set of exciting new developments from different optimization areas while at the same time providing a setting which will allow increased interaction among the participants. We aim to bring together researchers from both the theoretical and applied communities who do not usually get the chance to interact in the framework of a medium-scale event. INVITED TALKS: Distinguished guests will give one-hour long invited talks. Confirmed invited speakers include: John Dennis (Rice University) Michael Todd (Cornell University) Lieven Vandenberghe (UCLA) David Williamson (IBM Almaden) CONTRIBUTED TALKS: Contributions are solicited for presentation at the conference. Each accepted paper will be alloted a 25 minute talk. Presentation of recent results is encouraged. Authors wishing to speak at the conference should submit the abstract on which the talk will be based, in ASCII or LaTex source, to terlaky@mcmaster.ca by April 30, 2001. Please use "MOPTA" in the email subject line. All decisions on acceptance will be made after the closing day of April 30, 2001. A preliminary program will be available before the end of May. ORGANIZING COMMITTEE The Organizing Committee Tamas Terlaky, terlaky@mcmaster.ca (Chair) Stavros Kolliopoulos (CAS, McMaster University) Tom Luo (ECE, McMaster University) Henry Wolkowicz (Comb. Opt. University of Waterloo) ********************************************************* Prof. Tamas Terlaky Department of Computing and Software, Office: JHE/214 McMaster University Hamilton, Ontario, Canada, L8S 4L7 Phone: +1-905 525-9140 ext. 27780, FAX: +1-905 524-0340 Email: terlaky@cas.mcmaster.ca www: http://www.cas.mcmaster.ca/~terlaky ******************************************************** A new journal OPTIMIZATION AND ENGINEERING is established The first issue is on the WEB! Please look at the home page: http://ssor.twi.tudelft.nl/~terlaky/OPTE/opte.html http://www.cas.mcmaster.ca/~terlaky/OPTE/opte.html http://fmad-www.larc.nasa.gov/mdob/users/natalia/opte ********************************************************