(Exploiting Structure with Interior Point Method)

Very large optimization problems (LPs, QPs or NLPs) with millions of constraints and decision variables often display some special structure.

OOPS is the parallel interior point code that exploits any special structure in the Hessian and Jacobian matrices. The solver is an implementation of the primal-dual interior point method with multiple centrality correctors. The solver is implemented using object-oriented programming techniques. It solves linear (LP), quadratic (QP) and nonlinear (NLP) problems. The sequential code compares favourably with commercial packages and the parallel code shows perfect speed-ups.

OOPS has beed developed by Jacek Gondzio, Andreas Grothey and Robert Sarkissian.
In May 2005 OOPS was used to solve a QP problem with 109 variables.

The following papers describe OOPS and address its various applications:

A convenient access to OOPS is provided through Structured Modelling Language (SML) described in the paper: A baby demo version of OOPS is available with SML software.

Robert gave two talks about the design of OOPS:
  1. INFORMS, Cincinnati, May 3-5, 1999.
  2. SIAM, Atlanta, May 10-12, 1999.
Andreas talked about it at several occasions:
  1. NA, Dundee, 24-27 June, 2003.
  2. ISMP, Copenhagen, August 18-22, 2003.
  3. NA, Dundee, 28-30 June, 2005.
Jacek talked about it at several occasions:
  1. IFIP, Cambridge, July 12-16, 1999.
  2. APMOD, London, April 17-19, 2000.
  3. ISMP, Atlanta, August 7-11, 2000 (talk00.pdf).
  4. ISMP, Copenhagen, August 18-22, 2003 (talk03.pdf).

Benchmarks on PCs.

Comparison with Cplex 7.0 Solver (SUN workstation): (PDF file, 612 kB) (April, 2003).
Comparison with Cplex 9.1 Solver (Linux PC): (PDF file, 394 kB) (February, 2006).

Return to Jacek Gondzio's page.