The doctoral dissertations of the former Helsinki University of Technology (TKK) and Aalto University Schools of Technology (CHEM, ELEC, ENG, SCI) published in electronic format are available in the electronic publications archive of Aalto University - Aaltodoc.
|
|
|
Dissertation for the degree of Doctor of Science in Technology to be presented with due permission of the Department of Electrical and Communications Engineering for public examination and debate in Auditorium S4 at Helsinki University of Technology (Espoo, Finland) on the 10th of December, 2004, at 12 o'clock noon.
Overview in PDF format (ISBN 951-22-7405-1) [705 KB]
Dissertation is also available in print (ISBN 951-22-7398-5)
Optical networks pose a rich variety of new design and performance analysis problems. Typically, the static design problems belong to the field of combinatorial optimisation, whereas decision-making and performance analysis problems are best treated using appropriate stochastic models. This dissertation focuses on certain issues in resource allocation and performance evaluation of backbone wavelength-routed (WR) networks and metropolitan area optical burst switching (OBS) networks.
The first two parts of the thesis consider heuristic algorithms for the static routing and wavelength assignment (RWA) and logical topology design (LTD) problems that arise in the context of WR networks. In a static RWA problem, one is asked to establish a given set of lightpaths (or light trees) in an optical WR network with given constraints, where the objective often is to minimise the number of wavelength channels required. In LTD problem, the number of wavelength channels is given and one is asked to decide on the set of lightpaths so that, for instance, the mean sojourn time of packets travelling at the logical layer is minimised. In the thesis, several heuristic algorithms for both the RWA and LTD problems are described and numerical results are presented.
The third part of the thesis studies the dynamic control problem where connection requests, i.e. lightpath requests, arrive according to a certain traffic pattern and the task is to establish one lightpath at a time in the WR optical network so that the expected revenue is maximised or the expected cost is minimised. Typically, the goal of optimisation is to minimise some infinite time horizon cost function, such as the blocking probability. In this thesis, the dynamic RWA problem is studied in the framework of Markov decision processes (MDP). An algorithmic approach is proposed by which any given heuristic algorithm can be improved by applying the so-called first policy iteration (FPI) step of the MDP theory. Relative costs of states needed in FPI are estimated by on-the-fly simulations. The computational burden of the approach is alleviated by introducing the importance sampling (IS) technique with FPI, for which an adaptive algorithm is proposed for adjusting the optimal IS parameters at the same time as data are collected for the decision-making analysis.
The last part of the thesis considers OBS networks, which represent an intermediate step towards full optical packet switching networks. In OBS networks, the data are transferred using optical bursts consisting of several IP packets going to the same destination. On the route of the burst, temporary reservations are made only for the time during which the burst is transmitted. This thesis focuses on fairness issues in OBS networks. It is demonstrated that fairness can be improved by using fibre delay lines together with Just-Enough-Time protocol (JET). Furthermore, by choosing the routes in an appropriate way one can also reach a satisfactory level of fairness and, at the same time, lower the overall blocking probability. Possible scheduling policies for metropolitan area OBS ring networks are also studied.
This thesis consists of an overview and of the following 9 publications:
Keywords: optical networks, WDM, routing and wavelength assignment, Markov decision process, policy iteration, optical burst switching
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
© 2004 Helsinki University of Technology