We offer computer programs for mathematical optimization, aimed particularly at large-scale problems. They cover areas of

• nonlinear and semidefinite programming: our flagship PENNON will offer a unique opportunity to solve optimization problems with general smooth objective function and combination of standard nonlinear constraints with linear and bilinear matrix inequality constraints. Used only as NLP solver, it turnes into our former code PENNLP that is among the leading codes for general (smooth) large-scale nonlinear optimization and one of the fastest codes for (smooth) convex optimization.
  NEW: a free open source Matlab based code PENLAB can do (almost) everything that PENNON can. PENLAB was developed in cooperation with NAG Ltd.

• bilinear matrix inequalities:  PENBMI is the first available code that (locally) solves optimization problems with bilinear matrix inequality constraints. Together with the YALMIP parser it offers an extremely efficient tool for solving many problems of optimal control.

• linear semidefinite programming: the code PENSDP solves optimization problems with linear matrix inequality constraints. It is one of the most efficient codes available for large-scale sparse problems.

We further develop specialized codes for problems of structural optimization. PENFMO is a code for Free Material Optimization that helps to design optimal properties and distribution of (typically composite) material in elastic structures. In this area we also offer consultations and development of special-purpose codes. Please contact us for more details.

PENOPT solvers were developed by Michal Kocvara and Michael Stingl; if you have further questions, contact us.