Mathematical Programming

, Volume 87, Issue 2, pp 215–249

A primal-dual trust-region algorithm for non-convex nonlinear programming

Authors

  • Andrew R. Conn
    • IBM T.J. Watson Research Center, P.O.Box 218, Yorktown Heights, NY, USA, e-mail: arconn@watson.ibm.com
  • Nicholas I. M. Gould
    • Rutherford Appleton Laboratory, Computational Science and Engineering Departement, Chilton, Oxfordshire, England, e-mail: n.gould@rl.ac.uk
  • Dominique Orban
    • CERFACS, 42 Avenue Gaspard Coriolis, 31057 Toulouse Cedex 1, France, e-mail: Dominique.Orban@cerfacs.fr
  • Philippe L. Toint
    • Facultés Universitaires Notre-Dame de la Paix, 61, rue de Bruxelles, B-5000 Namur, Belgium, e-mail: Philippe.Toint@fundp.ac.be

DOI: 10.1007/s101070050112

Cite this article as:
Conn, A., Gould, N., Orban, D. et al. Math. Program. (2000) 87: 215. doi:10.1007/s101070050112

Abstract.

A new primal-dual algorithm is proposed for the minimization of non-convex objective functions subject to general inequality and linear equality constraints. The method uses a primal-dual trust-region model to ensure descent on a suitable merit function. Convergence is proved to second-order critical points from arbitrary starting points. Numerical results are presented for general quadratic programs.

Copyright information

© Springer-Verlag Berlin Heidelberg 2000