Numerical examples solved with SPENBAR – Modified penalty barrier method for large-scale nonlinear programming problems: SPENBAR implements a modified penalty algorithm for solving large-scale nonlinear programming problems. It uses only one-dimensional arrays and takes full advantage of the sparsity of the Jacobians of the inequality and equality constraints. Using a modified penalty method, the original problem is converted into a sequence of unconstrained minimization subproblems. The unconstrained optimization problems are solved by means of a truncated Newton method implemented in subroutine TN by Stephen Nash. The package has a number of parameters which fix the strategy of the optimization, as have been recommended by Andy Conn, Nick Gould and Philippe Toint. It is based on a version of an implementation given by M.G. Breitfeld - D.F. Shanno and K. Schittkowski.

Keywords for this software

Anything in here will be replaced on browsers that support the canvas element