Numerical experience with solving MPECs as NLPs

Roger Fletcher (fletchermaths.dundee.ac.uk)
Sven Leyffer (sleyffermaths.dundee.ac.uk)

Abstract: This paper describes numerical experience with solving MPECs as NLPs on a large collection of test problems. The key idea is to use off-the-shelf NLP solvers to tackle large instances of MPECs. It is shown that SQP methods are very well suited to solving MPECs and at present outperform Interior Point solvers both in terms of speed and reliability. All NLP solvers also compare very favourably to special MPEC solvers on tests published in the literature.

Keywords: MPEC, equilibrium constraints, nonlinear programming, SQP, interior point methods.

Category 1: Complementarity and Variational Inequalities

Category 2: Nonlinear Optimization (Constrained Nonlinear Optimization )

Category 3: Optimization Software and Modeling Systems (Optimization Software Benchmark )

Citation: Numerical Analysis Report NA/210, Department of Mathematics, University of Dundee, August 2002.

Download: [Compressed Postscript]

Entry Submitted: 08/20/2002
Entry Accepted: 08/20/2002
Entry Last Modified: 08/20/2002

Modify/Update this entry