Ladies Night for Women in Engineering Sciences

[more]

Mixed-Integer Nonlinear Optimization is the mother of all deterministic optimization models. As such, it has enormous modeling power, with applications in all kinds of areas like manufacturing and transportation logistics, design of water and gas networks, chemical engineering, portfolio optimization, etc. But of course in its full generality, it is foolhardy to consider algorithms and meaningful positive theoretical results. On the other hand, there are many well-known positive results for the linear case, so it is natural to seek to build up from the linear case, to get positive results (both theoretical and computational) for broader models. Natural extensions involve convexity and separability, and their relatives, and polynomials. A natural step in this direction involves attempting to exploit quadratic functions. I will survey some recent results in this direction --- both negative and positive complexity results and practical methods. [more]

Renewable energy supply chain optimization problems are characterized by a large number of options, significant amounts of uncertainty and multi-scale nature of decisions. This presentation examines these characteristics through practical examples and offers some promising research directions. The problem of large number of processing stages/options will be first introduced through a microalgae-based bio-refinery design problem. Superstructure based modeling and optimization will be presented as a tool to investigate the problem at a high level. Then, the presentation will move onto the issue of coupling between long-term planning decisions like capital investment and policy and shorter-term decisions like production capacity operation and logistics. This aspect manifests itself as a large number of decision variables and constraints complicating solution of the optimization. The optimization complexity gets greatly amplified when the issue of uncertainty is added to the problem. We will examine both two stage and multi-stage problems. Examples of biofuel processing supply chain and energy portfolio optimization for power generation will be used to bring out the essential features and complications. For solutions, stochastic programming and approximate dynamic programming will be introduced. [more]

The use of second order sliding modes, namely the generalized super-twisting algorithm or one of its variants, is illustrated for two applications in bioprocesses. Both the theoretical framework and some practical aspects are discussed. Two applications are considered: the design of a variable-gain super-twisting observer to robustly estimate the reaction rate in a bioreactor, and the proposal of a pseudo-super-twisting controller that maximizes the product output rate for a simplified model of a bioprocess. The case studies are the on-line estimation of the nitrogen quota in a microalgae bioreactor, and the maximization of biogas production during anaerobic wastewater treatment. [more]

In the last two decades the design of nonlinear model predictive control (NMPC) schemes has received widespread attention by theoreticians and practitioners in the field of systems and control. The main reasons for this interest are (a) that NMPC allows considering input and state constraints in a structured manner, and (b) that NMPC allows handling nonlinear systems with multiple inputs. NMPC is built upon the repeated solution of an optimal control problem and the partial application of optimal input trajectories. Often, stability of NMPC schemes is enforced by means of computationally intensive terminal constraints and end penalties. In this talk, we discuss the design of NMPC schemes based on turnpike properties. We show that these properties enable avoiding terminal constraints. We begin the talk with a formal introduction of turnpike properties of optimal control problems. It is worth to be mentioned that the concept of turnpike properties has received widespread attention in optimal control approaches to economic dynamic systems. However, it is surprising that turnpike properties have received only limited attention in the context of NMPC. In this presentation, we present results attempting to partially bridge this gap. We show that exact turnpikes allow establishing stability of NMPC controlled systems as well as recursive feasibility without any terminal constraints. We draw upon examples from different areas such as process control and biology to illustrate our results. [more]

[more]

Model Predictive Control (MPC) has become one of the most popular control techniques in the process industry mainly because of its ability to deal with multiple-input-multiple-output plants and with constraints. However, its performance can deteriorate in the presence of model uncertainties and disturbances. In the last years, the development of robust MPC techniques has been widely discussed, but these were rarely, if at all, applied in practice due to their conservativeness or their computational complexity. This talk presents multi-stage nonlinear model predictive control (multi-stage NMPC) as a promising non-conservative robust NMPC control scheme, which is applicable in real-time. The approach is based on the representation of the evolution of the uncertainty by a scenario tree. It leads to non-conservative robust control of the plant because it takes into account explicitly that new information (usually present in the form of measurements) will become available at future time steps and that the future control inputs can be adapted accordingly, acting as recourse variables. The approach is illustrated using a challenging industrial case-study, which shows that multi-stage NMPC is a promising strategy for the optimizing control of uncertain nonlinear systems subject to hard constraints. It is also shown that multi-stage NMPC performs better than standard NMPC or other robust NMPC approaches presented in the literature while still being implementable in real-time despite of the main challenge of the method: The size of the resulting optimization problem. Finally, it is illustrated that thanks to the flexibility of the presented approach, it is possible to integrate it with other existing methods to enhance its capabilities and performance. [more]