, Volume 52, Issue 2, pp 123–137 | Cite as

Node and edge relaxations of the Max-cut problem

  • S. Poljak
  • F. Rendl


We study an upper bound on the max-cut problem defined via a relaxation of the discrete problem to a continuous nonlinear convex problem, which can be solved efficiently. We demonstrate how far the approach can be pushed using advanced techniques from numerical linear algebra and nonsmooth optimization. Various classes of graphs with up to 50 000 nodes and up to four million edges are considered. Since the theoretical bound can be computed only with a certain precision in practice, we use duality between node- and edge-oriented relaxations to estimate the difference between the theoretical and the computed bounds.

AMS Subject Classification

90C06 90C27 

Key words

Max-cut problem semidefinite programming min-max eigenvalue problem 

Knoten- und Kanten-Relaxationen beim Max-Cut Problem


Es wird eine obere Schranke für das Max-Cut Problem untersucht, die sich aus einer Relaxation des diskreten Problems zu einem stetigen, nichtlinearen konvexen Problem ergibt. Die Relaxation ist polynomial lösbar. Es werden die Grenzen des Ansatzes unter dem Einsatz fortgeschrittener Methoden aus numerischer linearer Algebra und nichtglatter Optimierung untersucht. Verschiedene Graphenklassen mit bis zu 50 000 Knoten und 4 Millionen Kanten werden mit dem Ansatz behandelt. Da die theoretische obere Schranke in der Praxis nur mit einer gewissen Genauigkeit bestimmt werden kann, wird ein Dualitätsmodell zwischen knoten- und kantenorientierten Relaxationen verwendet, um den Unterschied zwischen der theoretischen und der berechneten Schranke abzuschätzen.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    Barahona, F.: The max-cut problem in graphs not contractible toK 5. Oper. Res. Lett.2, 107–111 (1983)Google Scholar
  2. [2]
    Barahona, F., Grötschel, M., Jünger, M., Reinelt, G.: An application of combinatorial optimization to statistical physics and circuit layout design. Oper. Res.36, 493–513 (1988).Google Scholar
  3. [3]
    Barahona, F., Jünger, M., Reinelt, G.: Experiments in quadratic 0–1 programming. Math. Programm.44, 127–137 (1989).Google Scholar
  4. [4]
    Barahona, F., Titan, H.: Max mean cuts and max cuts. In: Combinatorial optimization in science and technology (extended abstracts), DIMACS Report 91-18, Rutgers University, New Brunswick, New Jersey, pp. 30–45 1991.Google Scholar
  5. [5]
    Delorme, C., Poljak, S.: Laplacian eigenvalues and the maximum cut problem. Math. Programm.62, 557–574 (1993).Google Scholar
  6. [6]
    deSimone, C., Rinaldi, G.: A cutting plane algorithm for the max-cut problem. Technical Report Report 346, IASI Rome, 1992.Google Scholar
  7. [7]
    Deza, M., Laurent, M.: Applications of cut polyhedra. Technical Report LIENS-92-18, Ecole Normale Supérieure, Paris, 1992.Google Scholar
  8. [8]
    Hadlock, F. O.: Finding a maximum cut of a planar graph in polynomial time. SIAM J. Comput.4, 221–225 (1975).Google Scholar
  9. [9]
    Karp, R. M.: Reducibility among combinatorial problems. In: Complexity of computer computation (Miller, R. E., Thather, J. W., eds.), pp. 85–103. New York: Plenum Press 1972.Google Scholar
  10. [10]
    Laurent, M., Poljak, S.: The metric polytope. Proc. of IPCO 1992 (Balas, E., Cornuejols, G., Kannan, R., eds.), pp. 274–286, 1992.Google Scholar
  11. [11]
    Lengauer, T.: Combinatorial algorithms for integrated circuit layout. Chichester: J. Wiley 1990.Google Scholar
  12. [12]
    Mohar, B., Poljak, S.: Eigenvalues and the max-cut problem. Czech. Math. J.40, 343–352 (1990).Google Scholar
  13. [13]
    Orlova G. I., Dorfman, Y. G.: Finding the maximal cut in a graph. Eng. Cybernetics10, 502–504 (1972).Google Scholar
  14. [14]
    Pardalos, P. M., Rodgers, G. P.: Computational aspects of a branch and bound algorithm for quadratic zero-one programming. Computing45, 131–144 (1990).Google Scholar
  15. [15]
    Poljak, S.: Polyhedral and eigenvalue approximations of the max-cut problem. In: Sets, graphs and numbers, (Miklós, D., Halász, G., Lovász, L., Szönyi, T., eds.), pp. 568–581. Amsterdam: North-Holland 1992.Google Scholar
  16. [16]
    Poljak, S., Tuza, Zs.: On the expected relative error of the polyhedral approximation of the max-cut. Technical Report Report 92757-OR, Institut für Diskrete Mathematik, Universität Bonn, 1992.Google Scholar
  17. [17]
    Rendl, F., Poljak, S.: Nonpolyhedral relaxations of graph bisection problems. Technical Report, University of Technology Graz, 1992.Google Scholar
  18. [18]
    Rendl, F., Poljak, S.: Solving the max-cut problem using eigenvalues. Technical Report, University of Technology Graz, 1991.Google Scholar
  19. [19]
    Rendl, F., Vanderbei, R. J., Wolkowicz, H.: Interior-point methods for max-min eigenvalue problems. Technical Report SOR-93-15, Princeton University, 1993.Google Scholar
  20. [20]
    Schramm, H., Zowe, J.: A combination of the bundle approach and the trust region concept. In: Advances in mathematical optimization (Guddat, J., et al. eds.), pp. 196–209. Berlin: Akademie Verlag 1988.Google Scholar
  21. [21]
    Schramm, H., Zowe, J.: A version of the bundle idea for minimizing a nonsmooth function: conceptual idea, convergence analysis, numerical results. SIAM J. Optimization2, 121–152 (1992).Google Scholar
  22. [22]
    Scott, D. S.: Block Lanczos software for symmetric eigenvalue problems. Technical Report ORNL/CSD-48, Oak Ridge National Laboratory, 1979.Google Scholar
  23. [23]
    Simon, H.: Personal communication. NASA Ames Research Center, Moffett Field, CA, 1993.Google Scholar

Copyright information

© Springer-Verlag 1994

Authors and Affiliations

  • S. Poljak
    • 1
  • F. Rendl
    • 2
  1. 1.Department of Applied MathematicsCharles UniversityPraha 1Czech Republic
  2. 2.Institut für MathematikTechnische Universität GrazGrazAustria

Personalised recommendations