Global dynamic optimization with Hammerstein–Wiener models embedded

Abstract Hammerstein–Wiener models constitute a significant class of block-structured dynamic models, as they approximate process nonlinearities on the basis input–output data without requiring identification full nonlinear model. Optimization problems with embedded are nonconvex, and thus local optimization methods may obtain suboptimal solutions. In this work, we develop deterministic global strategy that exploits specific structure to extend existing theory systems linear dynamics. At first, discuss alternative formulations problem embedded, demonstrating careful selection variables can offer numerical advantages solution approach. Then, convex relaxations for proposed implementation aspects focusing control parametrization technique. Finally, apply our case studies comprising both offline online problems. The results confirm an improved computational performance approach over options not exploiting dynamics all considered examples. They also underline tractability when using few intervals in applications like model predictive control.