Incorporating Inequality Constraints in the Spectral Bundle Method

Semidefinite relaxations of quadratic 0–1 programming or graph partitioning problems are well known to be of high quality. How- ever, solving them by primal-dual interior point methods can take much time even for problems of moderate size. Recently we proposed a spectral bundle method that allows to compute, within reasonable time, ap- proximate solutions to structured, large equality constrained semidefinite programs if the trace of the primal matrix variable is fixed. The latter property holds for the aforementioned applications. We extend the spec- tral bundle method so that it can handle inequality constraints without seriously increasing computation time. This makes it possible to apply cutting plane algorithms to semidefinite relaxations of real world sized instances. We illustrate the efficacy of the approach by giving some pre- liminary computational results.

[1]  Jochem Zowe,et al.  A Version of the Bundle Idea for Minimizing a Nonsmooth Function: Conceptual Idea, Convergence Analysis, Numerical Results , 1992, SIAM J. Optim..

[2]  Alexander Shapiro,et al.  On Eigenvalue Optimization , 1995, SIAM J. Optim..

[3]  László Lovász,et al.  On the Shannon capacity of a graph , 1979, IEEE Trans. Inf. Theory.

[4]  David P. Williamson,et al.  Improved approximation algorithms for maximum cut and satisfiability problems using semidefinite programming , 1995, JACM.

[5]  C. Helmberg,et al.  Solving quadratic (0,1)-problems by semidefinite programs and cutting planes , 1998 .

[6]  L. Lovász,et al.  Polynomial Algorithms for Perfect Graphs , 1984 .

[7]  Krzysztof C. Kiwiel,et al.  Proximity control in bundle methods for convex nondifferentiable minimization , 1990, Math. Program..

[8]  Kim-Chuan Toh,et al.  On the Nesterov-Todd Direction in Semidefinite Programming , 1998, SIAM J. Optim..

[9]  Xiong Zhang,et al.  Solving Large-Scale Sparse Semidefinite Programs for Combinatorial Optimization , 1999, SIAM J. Optim..

[10]  Alexander Schrijver,et al.  Cones of Matrices and Set-Functions and 0-1 Optimization , 1991, SIAM J. Optim..

[11]  Franz Rendl,et al.  A Spectral Bundle Method for Semidefinite Programming , 1999, SIAM J. Optim..

[12]  Satissed Now Consider.  Improved Approximation Algorithms for Maximum Cut and Satissability Problems Using Semideenite Programming , 1997 .

[13]  Gene H. Golub,et al.  Matrix computations , 1983 .