PENOPT is a set of computer programs for mathematical optimization and related specialized optimization programs. Our goal is to develop a unified approach to problems of nonlinear programming (NLP) and (linear and nonlinear) semidefinite programming (SDP). The solvers are based on a generalized Augmented Lagrangian method combined with the Trust Region algorithm.

PENNLP - a solver for convex and nonconvex large scale nonlinear programming problems. It can be used as

• MATLAB function with PEN interface
• routine called from a C or Fortran program with data communication through user-supplied subroutines
• stand-alone program with AMPL interface

PENSDP is a solver for linear semidefinite programming problems. It is aimed at large-scale dense and sparse SDP problems. It can be used as

• stand-alone program with input file in the sparse SDPA input format
• MATLAB function with PEN or YALMIP interface (SDPA-to-PEN interface available)
• routine called from a C or Fortran program with data communication through parameters

PENBMI - a solver for optimization and feasibility problems with quadratic objective function and linear and bilinear matrix inequalities. It can be used as

• MATLAB function with PEN or YALMIP interface
• routine called from a C or Fortran program with data communication through parameters

PENFMO - a code for Free Material Optimization. Computes optimal material distribution/properties in an arbitrary two- and three-dimensional domain under single and multiple load conditions. Used as a stand-alone program.

PENNON - a solver for convex and nonconvex large scale nonlinear programming problems with vector and matrix variables and matrix inequality constraints. It can be used as

• MATLAB function with PEN or YALMIP interface
• stand-alone program with AMPL interface (only NLP problems)