Receding Horizon Optimal Control of Wiener Systems by Application of an Asymmetric Cost Function

Wiener models are an important class of nonlinear systems which well approximate many applications. Real time optimal control of Wiener models, for instance in the form of receding horizon optimal control, can be done using the nonlinear setup and corresponding nonlinear optimization tools. However, as this paper shows, under rather mild conditions on the static nonlinearity, it is possible to reformulate the optimal control problem as a linear problem with an asymmetric cost function, whose solution can be computed using a slack variable extension of the initial quadratic problem with a small additional computational cost. This paper shows the approach and the achievable performance at the example of the emission control of a large gas engine used in the U.S. pipeline network.