Journal of Combinatorial Optimization

, Volume 6, Issue 3, pp 237–270 | Cite as

Geometry of Semidefinite Max-Cut Relaxations via Matrix Ranks

  • Miguel F. Anjos
  • Henry Wolkowicz
Article

Abstract

Semidefinite programming (SDP) relaxations are proving to be a powerful tool for finding tight bounds for hard discrete optimization problems. This is especially true for one of the easier NP-hard problems, the Max-Cut problem (MC). The well-known SDP relaxation for Max-Cut, here denoted SDP1, can be derived by a first lifting into matrix space and has been shown to be excellent both in theory and in practice.

Recently the present authors have derived a new relaxation using a second lifting. This new relaxation, denoted SDP2, is strictly tighter than the relaxation obtained by adding all the triangle inequalities to the well-known relaxation. In this paper we present new results that further describe the remarkable tightness of this new relaxation. Let \(F_n \) denote the feasible set of SDP2 for the complete graph with n nodes, let Fn denote the appropriately defined projection of \(F_n \) into \(S^n \), the space of real symmetric n × n matrices, and let Cn denote the cut polytope in \(S^n \). Further let \(Y \in F_n \) be the matrix variable of the SDP2 relaxation and XFn be its projection. Then for the complete graph on 3 nodes, F3 = C3 holds. We prove that the rank of the matrix variable \(Y \in F_3 \) of SDP2 completely characterizes the dimension of the face of the cut polytope in which the corresponding matrix X lies. This shows explicitly the connection between the rank of the variable Y of the second lifting and the possible locations of the projected matrix X within C3. The results we prove for n = 3 cast further light on how SDP2 captures all the structure of C3, and furthermore they are stepping stones for studying the general case n ≥ 4. For this case, we show that the characterization of the vertices of the cut polytope via rank Y = 1 extends to all n ≥ 4. More interestingly, we show that the characterization of the one-dimensional faces via rank Y = 2 also holds for n ≥ 4. Furthermore, we prove that if rank Y = 2 for n ≥ 3, then a simple algorithm exhibits the two rank-one matrices (corresponding to cuts) which are the vertices of the one-dimensional face of the cut polytope where X lies.

semidefinite programming discrete optimization Lagrangian relaxation Max-Cut problem 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. M.F. Anjos, “New convex relaxations for the Maximum Cut and VLSI layout problems,” Ph.D. Thesis, University of Waterloo, 2001. Available online at http://etd.uwaterloo.ca/etd/manjos2001.pdf.Google Scholar
  2. M.F. Anjos and H. Wolkowicz, “Strengthened semidefinite relaxations via a second lifting for the Max-Cut problem,” Discrete Appl. Math., vol. 199, no. 1-2, pp. 79–106, 2002.Google Scholar
  3. F. Barahona, “The max-cut problem on graphs not contractible to K5,” Oper. Res. Lett., vol. 2, no. 3, pp. 107–111, 1983.Google Scholar
  4. F. Barahona, “On cuts and matchings in planar graphs,” Math. Programming, vol. 60, no. 1 (Ser. A), pp. 53–68, 1993.Google Scholar
  5. F. Barahona, M. Grötschel, M. J¨unger, and G. Reinelt, “An application of combinatorial optimization to statistical physics and circuit layout design,” Oper. Res., vol. 36, pp. 493–513, 1988.Google Scholar
  6. F. Barahona and A.R. Mahjoub, “On the cut polytope,” Math. Programming, vol. 36, no. 2, pp. 157–173, 1986.Google Scholar
  7. S.J. Benson, Y. Ye, and X. Zhang, “Solving large-scale sparse semidefinite programs for combinatorial optimization,” SIAM J. Optim., vol. 10, no. 2, pp. 443–461 (electronic), 2000.Google Scholar
  8. S. Burer and R.D.C. Monteiro, “A projected gradient algorithm for solving the maxcut SDP relaxation,” Optim. Meth. Softw., vol. 15, pp. 175–200, 2001.Google Scholar
  9. M.M. Deza and M. Laurent, Geometry of Cuts and Metrics, Springer-Verlag: Berlin, 1997.Google Scholar
  10. U. Feige and G. Schechtman, “On the optimality of the random hyper plane rounding technique for MAX CUT,” Technical report, Weizmann Institute, Rehovot, Israel, 2000.Google Scholar
  11. M.R. Garey and D.S. Johnson, Computers and Intractability: A Guide to the Theory of NP-Completeness, Freeman: San Francisco, 1979.Google Scholar
  12. M.X. Goemans, “Semi definite programming in combinatorial optimization,” Math. Programming, vol. 79, pp. 143–162, 1997.Google Scholar
  13. M.X. Goemans, “Semi definite programming and combinatorial optimization,” Documenta Mathematica, Extra Volume ICM 1998, pp. 657–666, 1998.Google Scholar
  14. M.X. Goemans and F. Rendl, “Combinatorial optimization,” in Handbook of Semidefinite Programming: Theory, Algorithms, and Applications, H. Wolkowicz, R. Saigal, and L. Vandenberghe (Eds.), Kluwer Academic Publishers: Boston, MA, 2000.Google Scholar
  15. M.X. Goemans and D.P. Williamson, “.878-approximation algorithms for MAX CUT and MAX 2SAT,” in ACM Symposium on Theory of Computing (STOC), 1994.Google Scholar
  16. M.X. Goemans and D.P. Williamson, “Improved approximation algorithms for maximum cut and satisfiability problems using semi definite programming,” J. Assoc. Comput. Mach., vol. 42, no. 6, pp. 1115–1145, 1995.Google Scholar
  17. J. Hastad, “Some optimal in approximability results,” in Proc. of the 29th ACM Symp. on Theory Comput., 1997.Google Scholar
  18. C. Helmberg, “An interior-point method for semi definite programming and max-cut bounds,” PhD Thesis, Graz University of Technology, Austria, 1994.Google Scholar
  19. C. Helmberg and F. Oustry, “Bundle methods to minimize the maximum eigenvalue function,” in Handbook of Semidefinite Programming: Theory, Algorithms, and Applications, H. Wolkowicz, R. Saigal, and L. Vandenberghe (Eds.), Kluwer Academic Publishers: Boston, MA, 2000.Google Scholar
  20. C. Helmberg, S. Poljak, F. Rendl, and H. Wolkowicz, “Combining semidefinite and polyhedral relaxations for integer programs,” in Integer Programming and Combinatorial Optimization (Copenhagen, 1995), Springer: Berlin, 1995, pp. 124–134.Google Scholar
  21. C. Helmberg and F. Rendl, “A spectral bundle method for semi definite programming,” SIAM J. Optim., vol. 10, no. 3, pp. 673–696, 2000.Google Scholar
  22. C. Helmberg, F. Rendl, R.J. Vanderbei, and H. Wolkowicz, “An interior-point method for semi definite programming,” SIAM J. Optim., vol. 6, no. 2, pp. 342–361, 1996.Google Scholar
  23. R.A. Horn and C.R. Johnson, Matrix Analysis, Cambridge University Press: Cambridge, 1990.Google Scholar
  24. R.M. Karp, “Reducibility among combinatorial problems,” in Complexity of Computer Computation, R.E. Miller and J.W. Thatcher (Eds.), Plenum Press: New York, 1972, pp. 85–103.Google Scholar
  25. J.B. Lasserre, “Optimality conditions and LMI relaxations for 0-1 programs,” LAAS research report 00099, LAAS-CNRS, Toulouse, France, 2000.Google Scholar
  26. M. Laurent, “Tighter linear and semi definite relaxations for max-cut based on the Lovász-Schrijver lift-and-project procedure,” SIAM J. Optim., vol. 12, pp. 345–375, 2001 (electronic publication).Google Scholar
  27. M. Laurent and S. Poljak, “On a positive semi definite relaxation of the cut polytope,” Linear Algebra Appl., vol. 223/224, pp. 439–461, 1995.Google Scholar
  28. M. Laurent and S. Poljak, “On the facial structure of the correlation matrices,” SIAM J. Matrix Anal. Appl., vol. 17, no. 3, pp. 530–547, 1996.Google Scholar
  29. T. Lengauer, Combinatorial Algorithms for Integrated Circuit Layout, JohnWiley & Sons Ltd.: Chichester, 1990.Google Scholar
  30. B. Mohar and S. Poljak, “Eigenvalues in combinatorial optimization,” in Combinatorial Graph-Theoretical Problems in Linear Algebra, IMA, vol. 50, Springer-Verlag: Berlin, 1993.Google Scholar
  31. Y.E. Nesterov, “Quality of semidefinite relaxation for non convex quadratic optimization,” Technical report, CORE, Universite Catholique de Louvain, Belgium, 1997.Google Scholar
  32. Y.E. Nesterov, H. Wolkowicz, and Y. Ye, “Semi definite programming relaxations of nonconvex quadratic optimization,” in Handbook of Semi definite Programming: Theory, Algorithms, and Applications, H. Wolkowicz, R. Saigal, and L. Vandenberghe (Eds.), Kluwer Academic Publishers: Boston, MA, 2000, p. 34.Google Scholar
  33. P. Parrilo, “Structured semi definite programs and semi algebraic geometry methods in robustness and optimization,” Ph.D. Thesis, Caltech, Pasadena, California, 2000.Google Scholar
  34. F. Rendl, “Semi definite programming and combinatorial optimization,” Appl. Numer. Math., vol. 29, pp. 255–281, 1999.Google Scholar
  35. H. Wolkowicz, R. Saigal, and L. Vandenberghe (Eds.), Handbook of Semi definite Programming: Theory, Algorithms, and Applications, Kluwer Academic Publishers: Boston, MA, 2000, pp. xxvi+654.Google Scholar
  36. Y. Ye, “Approximating quadratic programming with bound and quadratic constraints,” Math. Programming, vol. 84, pp. 219–226, 1999.Google Scholar

Copyright information

© Kluwer Academic Publishers 2002

Authors and Affiliations

  • Miguel F. Anjos
    • 1
  • Henry Wolkowicz
    • 1
  1. 1.Department of Combinatorics & OptimizationUniversity of WaterlooWaterlooCanada

Personalised recommendations