Examining Handovers in a Telecommunications Network Using Markov Chains and Dissimilarity Matrices
Abstract: A telecommunications network is divided into cells, which have radio properties to lessen interference. Users move between these cells with their equipment. If the equipment is actively used, it goes through a process called handover when it moves between cells, this creates sequences of visited cells. This thesis investigates these handovers and the corresponding sequences of visited cells. In this thesis there are two objectives related to the handovers between cells. The first is to determine if different types of sequences have different proportions of unwanted behaviour, the second is to develop a method to detect changes in the patterns of the handovers, between different time periods. For both objectives it is examined if the sequences of visited cells can be modelled as r-order Markov chains. For the first objective, it is examined if there are different proportions of unwanted behaviour for the r most recently visited cells, using a Markov chain approach. The sequences are also examined as a whole with a clustering method using dissimilarity matrices. For the second objective, it is first examined if it is possible to model the sequences of visited cells from different time periods as Markov chains and then perform a homogeneity test between them. After that it is examined if dissimilarity metrics could be used to detect changes between time periods, this is done using dissimilarity matrices. In the end it can be concluded that different types of sequences have different proportions of unwanted behaviour. Furthermore, it can be concluded that the approach of modelling the sequences as Markov chains in order to detect changes in handover behaviour between time periods, does not work. Finally, it is concluded that dissimilarity metrics could be used to detect changes between time periods, and additionally, some suitable dissimilarity metrics are presented.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)