|Cooperation:||TenneT TSO GmbH, ENSO NETZ GmbH, Venios GmbH, Energy Saxony e. V., MPI Magdeburg, TU Ilmenau|
|Support:||Federal Ministry of Education and Research (BMBF), Grant 05M18OCA, Projektträger DESY|
|Duration:||1.1.2018 - 31.12.2020|
|Description:||The stable operation of the various voltage layers of hierarchical electrical networks may profit from exploiting flexibilities provided by new power steering devices (FACTS) or flexibilities offered by power plants or batteries included in lower level networks. Starting at the smart grid level (TUI, JProf. Dr. Karl Worthmann) the project aims at providing reduced order models (MPI, Dr. Sara Grundel) of these flexibilities for each level so that exploiting these in optimal power flow methods (TUC/ADM, Prof. Dr. Christoph Helmberg; coordinator) allows to produce solutions with stable realizations over all levels (TUC/ACSD, Prof. Dr. Stefan Streif).|
|Duration:||1.7.2016 - 30.6.2019|
|Description:|| Highly volatile energy demands of households and new electric devices and volatile supply by renewables form a major challenge for the stability of electrical networks. Suppose data centers with significant demand for computing power are well spread in this network. Is it possible to support stability by moving computing requests between the data centers? What are algorithmic, what economic requirements for this to work?
This research topic is investigated in a young researchers group with three Ph. D. positions in Electrical Engineering, one postdoc in economics and one Ph. D. position in my group. We head this project and develop convex optimization techniques for coupling time expanded decision networks for the data centers and power plants with semidefinite relexations of the optimal power flow problem.
|Cooperation:||VOITH Engineering Services GmbH (now Leadec Holding BV & Co. KG)|
|Duration:||1.8.2015 - 31.7.2018|
|Description:|| For a (partially) automated production line each robot is given with a sequence of operations that it has to carry out periodically. Furthermore precedence relations and conflicts may be given between operations of distinct robots. Find minimal starting times for each operation so that precedences are observed, no conflicts arise and the periodic cycle time is minimized. This will be extended in several directions, e.g. to uncertain processing times (for some jobs carried out with human interaction) or to improve energy efficiency.
The project is part of "viRAL - validierte Inbetriebnahme von Roboteranlagen mit automatischer Logik- und Lageprüfung" headed by VOITH; a further partner is Fraunhofer IWU at Chemintz.
|Cooperation:||Deutsche Bahn AG|
|Support:||Federal Ministry of Education and Research (BMBF), Grant 05M13OCA, Projektträger DESY|
|Duration:||1.7.2013 - 30.6.2016|
|Description:|| The task is to find a feasible operational schedule for passenger and
freight trains within the German railway
network that minimizes either total or peak energy consumption.
In this, each train is given by a predefined routing and desired
stopping time windows. For feasibility station capacities and headway time
restrictions have to be observed. We will study possibilities to include the new energy aspect by convex robust approaches.
The project is part of the joint project E-Motion of Fraunhofer SCS Nürnberg, HSu Hamburg, TU Braunschweig, TU Chemnitz, Univ. Erlangen-Nürnberg, and ZIB (Konrad-Zuse-Zentrum Berlin).
|Cooperation:||Nokia Siemens Networks GmbH & Co. KG|
|Support:||Federal Ministry of Education and Research (BMBF), Grant 05M10OCB (previously 03MS616D), Projektträger DESY|
|Duration:||1.7.2010 - 30.6.2013 (extended to 31.12.2014)|
|Description:||How much capacity should be installed on the links in a backbone network of an internet provider so that all arising internet traffic can be routed through it with high probability? This is one of the central robust network design problems attacked within the joint effort ROBUKOM (robust communication networks) of TU Aachen, TU Berlin, TU Chemnitz, and ZIB (Konrad-Zuse-Zentrum Berlin). In this particular project we explore different possibilities offered by convex techniques for handling chance constraints for finding good integer capacity levels.|
|Cooperation:||eniPROD, TU Chemnitz Cluster of Excellence|
|Support:||State Ministry of Science and Arts (SMWK), based on EFRE funds of the European Union|
|Duration:||1.4.2009 - 28.2.2014|
|Description:||Under the leadership of the department of mechanical engineering, TU Chemnitz launched a common research effort directed at studying and improving energy efficiency in production. The focus of our work within this project are optimization models and methods in the sections Process Chains and Logistics and Factory Planning.|
|Cooperation:||Deutsche Bahn AG|
|Support:||Federal Ministry of Education and Research (BMBF), Grant 05M10OCD (previously 03MS640D), Projektträger DESY|
|Duration:||1.7.2010 - 30.6.2013 (extended to 31.7.2013)|
|Description:||Given a routing and stopping time windows for passenger and freight trains within the German railway network, the task is to find a feasible operational plan for all trains that observes station capacities and headway time restrictions. The previous project concentrated on possibilities to find such plans for long term simulations. The time frame for medium term planning requires faster methods and this needs new mathematical ideas. We explored possibilities offered by dynamic graph generation approaches and parallel methods for weakly coupled systems that can be decomposed by Lagrangian relaxation. This was part of a joint project KOSMOS of TU Braunschweig, TU Chemnitz, TU Dortmund, Univ. Erlangen-Nürnberg, and ZIB (Konrad-Zuse-Zentrum Berlin).|
|Cooperation:||Deutsche Bahn AG|
|Support:||Federal Ministry of Education and Research (BMBF), Grant 03HEPAG4, Projektträger BEO, Forschungszentrum Jülich|
|Duration:||1.7.2007 - 30.6.2010|
|Description:||For purposes of long term simulation the goal is to generate automatically, from given train data, a time table over a time slice of six hours for the entire German railway network, so that the resulting train schedule is conflict free. The main objectives are to manage the extreme size of the entire network and to develop approaches for increasing robustness of the operation plans with respect to delays or failures.|
|Cooperation:||Fiege eCom Gmbh & Co KG, Herlitz PBS AG|
|Support:||Federal Ministry of Education and Research (BMBF), Grant 03-HEM2B4, Projektträger BEO, Forschungszentrum Jülich|
|Duration:||1.4.2001 - 31.03.2004|
|Description:|| Our industrial partner operates several warehouses within
the same city. Some of these warehouses are too small to
hold sufficient stock of each product for the day; therefore,
goods are constantly transferred between warehouses by
a shuttle service consisting of several trucks. Part of
the demand is known in advance, part is stochastic and has
to be estimated from available statistical data.
The task is to find a schedule that distributes the goods between the warehouses by means of the trucks so that demand can be satisfied on time with high probability.
|Cooperation:||Herlitz PBS AG|
|Duration:||1997 - 2000|
|Description:|| For printing machines setup times account for a significant portion of the overall production time. They
depend heavily on the sequence of jobs. Based on a detailed model of the setup process a code has been developed that solves the following
scheduling problem: Given several non identical machines and a set of jobs, find an optimal
assignment of the jobs to the machines and a sequence on each machine so that the last job is
completed as early as possible.
The code is based on a polyhedral approach for solving an m-Cost ATSP. Current work concentrates on extending the code to further applications.