Designing a telecommunication network is a critical task for every company that has to rely on communication services. However, optimizing networks to satisfy the demands while keeping the expenses low is a difficult problem. Nowadays, network design is subject to a rapidly changing planning environment: New, improved network technologies are becoming available, the deregulated market leads to a variety of different tariffs and services. Furthermore, traffic patterns and characteristics in data communication change from day to day, and the overall traffic volumes show exponential growth. Therefore, the designer often only has very vague ideas about the environment in which the network design task takes place. Even if there is perfect knowledge about a present-day network design available, the rapidly changing environment makes it necessary to predict and take into account developments in the future. The thesis discusses the task of network design and formulates three types of telecommunication network design problems: "static", "incremental" and "evolutionary" (i.e. network planning over time). It gives an overview and an assessment of the degrees of freedom involved in the problems. A decision theoretic framework for network design which is capable of coping with all types of problems is proposed. This decision framework structures the design problem in an appropriate way, providing the designer with uniform guidelines on how to proceed, given an arbitrary planning environment. Furthermore, it is capable of incorporating incomplete knowledge. As a result of the theoretical concepts, an object-oriented application framework for telecommunication network design is presented that supports code and design reuse on different levels of abstraction. The last part of the thesis exemplarily applies the theoretical concepts by the use of a meta-heuristic and in the context of a real world problem, the network of the German Forschungsnetz-Verein (WiN).     «

Designing a telecommunication network is a critical task for every company that has to rely on communication services. However, optimizing networks to satisfy the demands while keeping the expenses low is a difficult problem. Nowadays, network design is subject to a rapidly changing planning environment: New, improved network technologies are becoming available, the deregulated market leads to a variety of different tariffs and services. Furthermore, traffic patterns and characteristics in data...     »

Die Planung einer unternehmensweiten Kommunikationsinfrastruktur ist eine immer wichtiger werdende Aufgabe. Ein optimales Telekommunikationsnetz mit minimalen Kosten zu finden ist jedoch anerkanntermaßen schwierig. Die Geschwindigkeit in der neue Technologien verfügbar werden und die Wachstumsraten des Datenverkehrs (Inter-/Intranet) stellen neue Anforderungen an die Planung in den Mittelpunkt: evolutionäre Planungsaspekte (d.h. Planung über die Zeit) und Planung mit unsicherem Wissen. Beide Aspekte werden in der Arbeit behandelt. Die theoretischen Ergebnisse werden in einem objekt-orientierten Software Framework modelliert. Die praktische Umsetzung wird exemplarisch an der Planungsaufgabe des deutschen Wissenschaftsnetzes gezeigt.     «

Die Planung einer unternehmensweiten Kommunikationsinfrastruktur ist eine immer wichtiger werdende Aufgabe. Ein optimales Telekommunikationsnetz mit minimalen Kosten zu finden ist jedoch anerkanntermaßen schwierig. Die Geschwindigkeit in der neue Technologien verfügbar werden und die Wachstumsraten des Datenverkehrs (Inter-/Intranet) stellen neue Anforderungen an die Planung in den Mittelpunkt: evolutionäre Planungsaspekte (d.h. Planung über die Zeit) und Planung mit unsicherem Wissen. Beide Asp...     »