The research project VINO deals with mathematical models and solution methods for problems that arise from the dynamic of modern ICT-networks. Regarding the requirements of the operators and users of such networks particular focus lies on the dynamic optimization of Flexgrid optical transport networks as well as on the embedding of virtual networks into physical substrate networks. The goal of the project is the development of new mathematical approaches, which provide solutions of both problem classes under consideration of temporal changes of the data and requests. Furthermore, the aim is the development and implementation of algorithms, which build the basis for practically applicable tools.
The project is divided into four subprojects, each focusing on a particular aspect of the whole project. These are 1) multi-period-optimization under consideration of temporal, non-stochastic changes, 2) static, detailed models for the generation of bounds for comparison purposes, 3) pro-active long-term planning models under uncertainties and 4) online-optimization and fast reconfiguration algorithms. An important aspect throughout the project is the active exchange between the subprojects and the corporation with our industrial partners.
The results of the research project will be made available to the involved companies. This includes the developed algorithms and software prototypes as well as relevant know-how gained in the project, giving the partners an advantage in the competition.
Project 1: Modellaggregation für zeitdiskrete und ereignisbasierte Modelle
(Model aggregation for time discretized and event based models)
This subproject aims at the development of efficient mathematical approaches for dynamic temporal aspects in mid-term and long-term optimization of networks. Main aspects are reduction and adaption techniques for time discretized and event based linear mixed-integer optimization models, which include both, approaches for a-priori aggregation as well as adaptive aggregation during the solution process. This allows to improve the practical tractability and the temporal resolution of these models so that practically relevant solutions can be obtain in reasonable time.
(Proactive optimization model for networks with reliability requirements)
The subproject aims at providing solutions to the VNE and Flexgrid problems in cases of data uncertainty. We assume that the nominal data taken as input for the two problems might differ from the real data due to, e.g., measurement errors or unpredictable changes in what was considered as a given. In this setting, the focus is on defining so-called uncertainty sets, representing (explicitly or implicitly) a set of realistic scenarios, e.g., of realizations of the uncertain data. Mathematical models and optimization algorithms are then designed so to produce solutions that are robust under uncertainty, i.e., feasible with respect to all the realizations described in the uncertainty sets. Read more ..
Project 3: Detaillierte und skalierbare mathematische Modelle für das VNE-Problem
(Detailed and scalable mathematical models for the VNE problem)
The aim of this subproject is to provide solutions for network embedding problems that obey as many detailed and technical side constraints as possible. This can be seen as a "snapshot" version of the dynamic VNE problems investigated by the partner projects. A comparison with such "ideal" solutions allows to estimate the quality of the solutions produced under the various prerequisites considered in the partner projects and a thorough analysis gives hints as to which aspects of the problem are particularly tricky to incorporate.
Project 4: Online-Algorithmen und Rekonfiguration im laufenden Netzbetrieb
Chemnitz University of Technology
Prof. Dr.-Ing. Thomas Bauschert
ADVA Optical Networking SE
Contact: Dr. Achim Autenrieth
Deutsche Telekom AG
Contact: Dr. Stefan Schmid, Dr. Fritz-Joachim Westphal
Nokia Siemens Networks Management International GmbH
Contact: Ernst-Dieter Schmidt
||| Optimal offline virtual network embedding with rent-at-bulk aspects ( ), In CoRR, volume abs/1501.07887, 2015. |
|||On the generation of cutting planes which maximize the bound improvement ( ), In International Symposium on Experimental Algorithms (SEA), 2015. |
||| Network design with compression: complexity and algorithms ( ), In INFORMS Computing Society Conference (INFORMS ICS), 2015. |
|||On the computational complexity of the virtual network embedding problem ( ), In International Network Optimization Conference (INOC), 2015. |
||| Virtual network embedding under uncertainty: exact and heuristic approaches ( ), In Design of Reliable Communication Networks (DRCN), 2015. |
|||Online Routing and Spectrum Assignment in Flexgrid Optical Networks ( ), In International Conference on Transparent Optical Networks, 2015. |
||| Mobile core network virtualization: A model for combined virtual core network function placement and topology optimization ( ), In Proceedings of the 1st IEEE Conference on Network Softwarization, NetSoft 2015, London, United Kingdom, April 13-17, 2015, 2015. |
||| Maximum throughput network routing subject to fair flow allocation ( ), In Combinatorial Optimization, LNCS, volume 8596, 2014. |