A new paradigm arising in feedback control is the incorporation of detailed DAE/PDAE simulation models in on-line environments. The main advantage of doing this is the ability to optimize, in real-time, the process operating conditions as external economic factors evolve in time (e.g. energy prices). On the other hand, this new generation of optimization-based controllers requires new and powerful nonlinear programming (NLP) techniques as fast solutions are critical to preserve the stability of the process. In this seminar, we discuss how warm-starting, parametric optimization and linear algebra techniques can be used to minimize the on-line solution times. In addition, we present different techniques to synchronize the dynamics of the process with the solution time of the NLP problems.