MISQPOA: A Fortran Subroutine for Mixed-Integer Nonlinear
Optimization by Outer Approximation
- User's Guide
T. Lehmann, K. Schittkowski, Report, Department of Computer Science, University of Bayreuth (2009)
Abstract:
The Fortran subroutine MISQPOA solves
mixed-integer nonlinear programming problems. It combines an outer approximation
method with a sequential quadratic programming trust region algorithm called
MISQP. If problem functions are relaxable
and analytical derivatives can be provided also for integer variables,
MISQPOA guarantees convergence towards a global
optimal solution for convex problems. Additional safeguards allow to apply
MISQPOA also to non-convex and non-relaxable
nonlinear mixed-integer programs, but without guaranteeing global optimality. In
the non-relaxable case, derivatives are approximated at neighbored grid pints.
To download the report, click here: misqpoa.pdf
(Acrobat Reader version)