An optimal control approach to nonlinear mixed integer programming problems

Motivated by the recent developments of the Control Parametrization Enhancing Technique (CPET), a novel method for solving a general class of nonlinear mixed integer programming problems is introduced in this paper. By imposing appropriate dynamics as well as as a set of statistical variance type of functional constraints, a problem with mixed integer decision variables is first transformed into a discrete-valued optimal control problem, and then transformed, by applying CPET, into a standard optimization problem involving only continuous values.