The control of large-scale networks using distributed computing platforms represents a key challenge for tackling a range of emerging control problems in important application areas such as power or transportation systems. Distributed Model Predictive Control (DMPC) methods aim at providing high-performance control under both system constraints and under limited locally available information about the network. A main difficulty is imposed by topology changes during closed-loop control, i.e. sub-systems joining or leaving the network, which generally invalidates guarantees on stability and constraint satisfaction of the current controller, while a re-design of all controllers in the network is infeasible.
This talk will present a new approach for plug & play distributed predictive control as well as for its high-speed implementation. First, a new concept of distributed invariance will be presented that is based on time-varying local sets, restoring some of the benefits of centralized MPC formulations, while allowing for a completely distributed synthesis and online control. We will then introduce a plug & play protocol for ensuring safety when topologies change online. The approach will be demonstrated for integrated electric vehicle charging with constrained voltage control in distribution grids. Finally, the talk will propose a high-speed solver based on a splitting method together with primal complexity bounds for certifying the computational procedure of the proposed plug & play DMPC scheme.
Melanie Zeilinger is a Postdoctoral Researcher and Marie Curie fellow in a joint program with the Max-Planck Institute for Intelligent Systems, Tuebingen, Germany and the University of California at Berkeley, USA. From 2011-2012 she was a postdoctoral fellow in the Automatic Control Laboratory at the École Polytechnique Fédérale de Lausanne (EPFL), Switzerland. In 2011, she received the Dr.sc. degree with honors in Electrical Engineering from ETH Zurich, Switzerland and in 2006, the diploma in Engineering Cybernetics from the University of Stuttgart, Germany. She conducted her diploma thesis research at the Department of Chemical Engineering, University of California at Santa Barbara, USA, in 2005-2006. Her general research interests are centered around real-time and distributed control and optimization as well as safe learning-based control, with an application focus on green energy-efficient technologies.