Showing papers by "Hans Georg Bock published in 2018"

Numerical solution of optimal control problems with explicit and implicit switches

Abstract: In this article, we present a unified framework for the numerical solution of optimal control problems (OCPs) constrained by ordinary differential equations with both implicit and explicit switches...

Sigmapoint Approach for Robust Optimization of Nonlinear Dynamic Systems

Abstract: Mathematical models describing dynamic processes contain parametric uncertainties. Robust model-based optimization thus becomes a challenging task in process engineering. Current approaches either require high computational effort or they make use of oversimplified approximations that do not capture changes in the solution structure due to nonlinear effects of the uncertain parameters on the states of the process. In this paper we propose an improved optimization approach that uses sigmapoints to characterize the space of uncertain parameters. Propagating sigmapoints through the process model and directly using them in the optimization problem allows to capture relevant nonlinearities for the uncertain parameters. Main advantages of this simple yet elegant approach are the relatively low computational burden and the independence from the optimizer, as no further derivatives are needed. The approach is applied to two examples from process engineering, a batch distillation and a semibatch reactor.