[pt] Este trabalho apresenta um modelo de tarifação dinâmica cujo objetivo é determinar qual o preço que deve ser cobrado por uma chamada telefônica, originada em uma estação móvel, de forma a maximizar a receita da empresa operadora do serviço móvel celular, garantindo os valores máximos aceitáveis para a probabilidade de bloqueio de uma chamada e para a probabilidade de queda de ligações devido ao handoff. / [en] This work presents a dynamic pricing model whose objective is to determinate the price that should be charged for a telephone call originated in a mobile phone in order to maximize the telephone enterprise revenue as long as it guarantees the maximum acceptable values for the probability of blocking of originating calls and for calls that are requesting a handoff. / [es] Este trabajo presenta un modelo de tarifación dinámica cuyo objetivo es determinar cuál es el precio que deve ser cobrado por una llamada telefónica, originada en una estación móvil, de tal forma que maximize la receta de la empresa operadora del servicio móvil celular, garantizando los valores máximos aceptables para la probabilidad de bloqueo de una llamada y para la probabilidad de caída de conexión debido al handoff.

Η παρούσα εργασία διερευνά την παράμετρο της συμφόρησης της επικοινωνιακής κίνησης στα κυψελωτά δίκτυα κινητής τηλεφωνίας GSM. Αρχικά γίνεται μια εισαγωγή στα δίκτυα κινητής τηλεφωνίας ξεκινώντας από τα συστήματα 1ης γενιάς και καταλήγοντας τα σημερινά συστήματα 4ης γενιάς. Στην συνέχεια, παρουσιάζεται η δομή του κυψελωτού δικτύου κινητής τηλεφωνίας GSM δίνοντας έμφαση στην αρχιτεκτονική του, στις διεπαφές και το πρωτόκολλα που χρησιμοποιεί καθώς και στις βασικές λειτουργίες που εκτελεί. Σε μεγαλύτερο βάθος, αναλύονται οι λειτουργίες και τα πρωτόκολλα του HLR κόμβου, ο οποίος αποτελεί και ένα από τα πιο κεντρικά μέρη του δικτύου GSM. Μετά μελετάται η παράμετρος της συμφόρησης στα δίκτυα GSM γενικά δίνοντας τον ορισμό της έννοιας της συμφόρησης, παρουσιάζοντας τα διαφορετικά είδη συμφόρησης και τις αιτίες που την δημιουργούν και καταλήγοντας στους ποικίλους τρόπους αντιμετώπισης της. Το βασικό κομμάτι της παρούσας εργασίας ασχολείται με την διερεύνηση της παραμέτρου της τηλεπικοινωνιακής συμφόρησης μέσα στο HLR κόμβο ενός κυψελωτού δικτύου κινητής τηλεφωνίας GSM, και συγκεκριμένα την συμφόρηση της κεντρικής μονάδας επεξεργασίας CPU του κόμβου αυτού. Περιγράφεται η περίπτωση του ενσωματωμένου HLR κόμβου (Integrated MSC/HLR) ενώ αναλύεται, σε πιο μεγάλη λεπτομέρεια, η περίπτωση του αυτόνομου HLR κόμβου (Stand-alone HLR). Ακολουθεί η παρουσίαση των αιτιών συμφόρησης της CPU ενός αυτόνομου HLR κόμβου και δίνεται ιδιαίτερη προσοχή στην διαδικασία του Location Updating και στην συμφόρηση που συμβαίνει κατά την διάρκεια αυτής της διαδικασίας. Γίνεται μια σύντομη αναφορά σε διάφορους τρόπους αντιμετώπισης αυτού του είδους της συμφόρησης και ύστερα προτείνεται ένας ολοκληρωμένος αλγόριθμος αποφυγής της εν λόγω συμφόρησης. Παρουσιάζεται με λεπτομέρεια το κάθε κομμάτι του αλγορίθμου αυτού και κατασκευάζεται ένας κώδικας προσομοίωσης του. Με βάση αυτών τον κώδικα, πραγματοποιούνται διάφορες μετρήσεις και πειράματα, παρουσιάζονται τα διάφορα αποτελέσματα και τελικά ο αλγόριθμος κρίνεται για την τη καταλληλότητα, την απόδοση, την προσαρμοστικότητα και την αξιοπιστία του. / This paper investigates the parameter of congestion of communication traffic in cellular mobile networks GSM. Initially, an introduction in the mobile networks is presented starting from the first generation systems and ending to the current fourth-generation systems. Then, the structure of cellular mobile GSM network is explored with an emphasis on architecture, interfaces and protocols used and the basic functions it performs. The functions and protocols of the HLR node, which is also one of the most central parts of the network GSM, is analyzed in more detail. Additionally, the parameter of the congestion in GSM networks, in general, is considered giving the definition of congestion, showing the different types and the causes that lead to its appearance and ending to the several ways that can help to prevent its creation or to suppress it. The main part of this study deals with the exploration of the parameter of congestion of telecommunication traffic in the HLR node in a cellular mobile network GSM, namely the congestion of central processing unit CPU of this node. It describes the case of integrated HLR node (Integrated MSC / HLR) and analyzes, in greater detail, the case of autonomous HLR node (Stand-alone HLR). The presentation of the causes of CPU bottlenecks in an autonomous HLR node follows with a particular attention to the Location Updating procedure and the congestion that occurs during this process. A brief reference to the various ways of dealing with this kind of congestion is presented and then an integrated algorithm that aims to avoid the congestion is proposed. Every piece of this algorithm is analyzed in great detail and a simulation code is constructed . Based on this code, various measurements and experiments are carried out , the different results are studied and evaluated and the algorithm is finally judged for its appropriateness, efficiency, adaptability and reliability.

Context. With the increased usage of mobile devices and internet, the cellular network traffic has increased tremendously. This increase in network traffic has led to increased occurrences of communication failures among the network nodes. Each communication failure among the nodes is defined as a bad event and occurrence of one such bad event acts as a source of origin for several consecutive bad events. These bad events as a whole may eventually lead to node failures (not being able to respond to any data requests). But it requires a lot of human effort and cost to be invested in by the telecom companies to implement workarounds for these node failures. So, there is a need to prevent node failures from happening. This can be done by classifying the traffic patterns between nodes in the network, identify bad events in them and deliver the verdict immediately after their detection. Objectives. Through this study, we aim to find the best suitable machine learning algorithm which can efficiently classify the traffic patterns of SGSN-MME (SGSN (Serving GPRS (General Packet Radio Service) Support node) and MME (Mobility Management Entity). SGSN-MME is a network management tool designed to support the functionalities of two nodes namely SGSN and MME. We do this by evaluating the classification performance of four machine learning classification algorithms, namely Support vector machines (SVMs), Naïve Bayes, Decision trees and Random forests, on the traffic patterns of SGSN and MME. The selected classification algorithm will be developed in such a way that, whenever it detects a bad event, it notifies the user about it by prompting a message saying, “Something bad is happening”. Methods. We have conducted an experiment for evaluating the classification performance of our four chosen classification algorithms on the dataset provided by Ericsson AB, Gothenburg. The experimental dataset is a combination of three logs, one of which represents the traffic patterns in real network and the other two logs contain synthetic traffic patterns that are generated manually. The dataset is unlabeled with 720 data instances and 4019 attributes in it. K-means clustering is performed for dividing the data instances into groups and thereby proceed with labeling them accordingly into good and bad events. Also, since the number of attributes in the experimental dataset are more than the number of instances, feature selection is performed for selecting the subset of relevant attributes which best represents the whole data. All the chosen classification algorithms are trained and tested with ten-fold cross validation sets using the selected subset of attributes and the obtained performance measures like classification accuracy, F1 score and training time are analyzed and compared for selecting the best suitable one among them. Finally, the chosen algorithm is tested on unlabeled real data and the performance measures are analyzed in order to check if is able to detect the bad events correctly or not. Results. Experimental results showed that Random forests outperformed Support vector machines, Naïve Bayes and Decision trees with an average classification accuracy of 99.72% and average F1 score of 99.6, when classification accuracy and F1 score are considered. On the other hand, Naive Bayes outperformed Support vector machines, Decision trees and Random forests with an average training time of 0.010 seconds, when training time is considered. Also, the classification accuracy and F1 score of Random forests on unlabeled data are found to be 100% and 100 respectively. Conclusions. Since our study focuses on classifying the traffic patterns of SGSN-MME more accurately, classification accuracy and F1 score are of highest importance than the training time of algorithm. Therefore, based on experimental results, we conclude that Random forests is the best suitable machine learning algorithm for classifying the traffic patterns of SGSN -MME. However, Naive Bayes can be also used if classification has to be performed in the least time possible and with moderate accuracy (around 70%).

[pt] Esta tese apresenta novos algoritmos para controle de admissão de usuários em redes móveis celulares. É utilizada a técnica de reserva de recursos, também conhecida por uso de canais de guarda, para atingir os graus de qualidade de serviço desejados para cada tipo de usuário. São propostos algoritmos dinâmicos, capazes de se adaptar ao perfil de tráfego presente na rede e que possuem diferentes filosofias de projeto. Inicialmente, foi considerado o caso de uma classe que resulta em dois tipos de usuários: chamadas novas e chamadas em handoff. Os algoritmos propostos são testados em condições de tráfego representadas por diversas distribuições para o tempo de permanência do usuário na célula. Foi desenvolvido um novo simulador em linguagem C que é capaz de verificar o desempenho dos algoritmos propostos. Resultados analíticos para desempenho dos algoritmos de uma classe e um número fixo de recursos reservados são apresentados empregando uma modelagem por Cadeia de Markov. Foi desenvolvido um método que permite calcular a intensidade de tráfego máxima a qual o sistema pode ser submetido, e a quantidade de recursos a ser reservada assumindo que o objetivo é maximizar a utilização do sistema atendendo os valores de qualidade de serviço estabelecidos, no caso de tempo de retenção do recurso de rádio modelado por uma chamada com distribuição exponencial. Foi proposto um algoritmo simples, dinâmico e distribuído, baseado em medidas em tempo real, cuja meta é acompanhar a curva ótima de número de recursos reservados. Posteriormente, os resultados analíticos empregando Cadeia de Markov são generalizados para M classes. Alguns dos algoritmos definidos para o caso de uma classe são estendidos para o caso de duas classes e seu desempenho é avaliado, utilizando o simulador desenvolvido neste trabalho. O método para calcular a intensidade máxima de recursos que o sistema comporta, sem violar os requisitos de qualidade de serviço, é estendido para o caso de duas classes. Finalmente, são definidos parâmetros que permitem comparar o desempenho dos algoritmos com 2M classes, considerando uma distribuição genérica para o tempo de permanência do usuário na célula. / [en] This thesis presents new algorithms for Channel Admission Control in wireless communications systems. We investigate techniques based in resource reservation, also known as guard channel, to achieve the quality of service desired for each class of users. We propose dynamic schemes based in the cell traffic. Each algorithm has a different goal, some try to minimize the probability of handoff fail, others try to maximize the traffic intensity when the limit imposed by QoS is being approached. First, we considered one class (M = 1) divided in two classes: new users and handoff users. In order to test the new schemes we developed a simulator in C that uses different distributions for the dwell-time. During the simulation, the measures of channel solicitations and the result of their allocation are used to decide whether new calls will be admitted. We also obtained analytic results using a Markov Chain model. We developed a method to calculate the maximum traffic intensity that the system supports without violating the established quality of service constraints, assuming one class of users and the dwell-time modelled by a exponential distribution. This method allows to identify the maximum traffic intensity supported by the system and also the exact number of resources to be reserved for each value of traffic intensity. We proposed a new, dynamic and distributed algorithm based on real time measures which targets to follow the optimum number of reserved curve obtained from our procedure. We generalized the analytic results using M-dimensional Markov Chains to 2M classes of users. Some of the algorithms defined to two classes (M = 1) were extended to the case of four classes (M = 2) and their performances are evaluated using the simulator developed in this work. The method to evaluate the maximum intensity of traffic within the limits of QoS is also extended to the case of four classes. Finally we define new parameters that allow the performance comparison among 2M class algorithms, considering any dwell- time distribution.

Στα κυψελωτά δίκτυα κινητών επικοινωνιών, η διαρκής κίνηση των κινητών συσκευών δημιουργεί την ανάγκη ύπαρξης μηχανισμών οι οποίοι θα διασφαλίζουν το αδιάλειπτο της επικοινωνίας «εν κινήσει». Αυτό ακριβώς επιτυγχάνεται με την εφαρμογή μηχανισμών μεταπομπής, οι οποίοι, με τρόπο διάφανο προς την κινητή συσκευή, συνδέονται δυναμικά με το καταλληλότερο σημείο εκπομπής τηλεπικοινωνιακού σήματος (Σταθμός Βάσης) που εκπέμπει στην ευρύτερη περιοχή. Στην παρούσα μελέτη παρουσιάζονται και αναλύονται οι υπάρχουσες κατηγορίες μεταπομπής, ενώ δίνεται ιδιαίτερη έμφαση σε αυτές που είναι εφαρμόσιμες στα δίκτυα τρίτης γενιάς (3G). Μία από τις σημαντικότερες κατηγοριοποιήσεις των τύπων μεταπομπής είναι αυτή σε σκληρού τύπου (hard) και μαλακού τύπου (soft ή softer). Σημαντικό πλεονέκτημα του μαλακού τύπου μεταπομπής είναι η διασφάλιση της ανεξαρτησίας καναλιών (channel diversity) η οποία τελικά οδηγεί σε αλγορίθμους μεταπομπής που καταναλώνουν μικρότερες ποσότητες ενέργειας. Από την άλλη στις μεταπομπές μαλακού τύπου γίνεται χρήση περισσότερων πόρων του δικτύου και υπάρχει μεγαλύτερη πολυπλοκότητα. Στην παρούσα μελέτη μελετώνται οι τεχνικές οι οποίες χρησιμοποιούνται για να ληφθεί η απόφαση εκτέλεσης μεταπομπής. Βασικές κατηγορίες αλγορίθμων απόφασης μεταπομπής είναι αυτοί που στηρίζονται στη στάθμη της ισχύος του λαμβανόμενου σήματος, αυτοί που στηρίζονται σε fuzzy logic και οι αλγόριθμοι προτεραιότητας. Στο πλαίσιο της παρούσας εργασίας μελετάται και αναλύεται κριτικά η σχετική βιβλιογραφική έρευνα που έχει διενεργηθεί και εξακολουθεί να διεξάγεται σχετικά με την εφαρμογή αλγορίθμων μεταπομπής, με έμφαση στις τεχνικές μεταπομπής που βρίσκουν εφαρμογή σε συστήματα CDMA (Code Division Multiple Access). Επιπλέον, υλοποιείται και προσομοιάζεται, ένα μοντέλο το οποίο συνδυάζει τεχνικές «Γκρι Προβλέψεων» (Grey Prediction) με κλασσικούς μηχανισμούς μεταπομπής με στόχο τη βελτιστοποίηση του πλήθους των εκτελούμενων μεταπομπών, ελαχιστοποιώντας ταυτόχρονα την πιθανότητα απόρριψης κλήσεων (call blocking probability). Η ταυτόχρονη χρήση και αξιοποίηση και των τριών εργαλείων (Matlab, Simulink και Stateflow) απλοποιεί την υλοποίηση και ταυτόχρονα διευκολύνει την αξιολόγηση των παραγόμενων αποτελεσμάτων. Ο βασικός δείκτης επίδοσης με τον οποίο αξιολογείται ο αλγόριθμος που υλοποιείται, είναι το πλήθος των μεταπομπών που εκτελούνται, ενώ οι εκτελεσθείσες προσομοιώσεις στηρίζονται σε τιμές των παραμέτρων που έχουν προταθεί στο (Saeed Changiz Rezaei, Hossein Khalaj, 2005), συνδυασμένες με διάφορες τιμές Υστέρησης (Hysteresis). Προκύπτει ότι καθώς αυξάνεται η τιμή της Υστέρησης, ο αριθμός των μεταπομπών που εκτελούνται μειώνεται, όπως άλλωστε αναμενόταν. Τέλος, θα πρέπει να επισημανθεί ότι, η πολύ μεγάλη συχνότητα δειγματοληψιών του εν λόγω μοντέλου σε συνδυασμό με την ομαλότητα κίνησης της κινητής συσκευής και την καλή προβλεπτική ικανότητα του Αλγορίθμου Grey Prediction, οδηγεί στον υπολογισμό μιας πολύ ομαλής ακολουθίας διαδοχικών τιμών ισχύος των λαμβανόμενων σημάτων και τελικά σε ελαχιστοποίηση του αριθμού των μεταπομπών που διενεργούνται. / In cellular mobile networks, the continuous movement of mobile devices creates the need for mechanisms that are necessary to ensure continuity of communication. This is exactly what is achieved by applying handover mechanisms, which connect the mobile device, dynamically and transparently with appropriate emission point (base station) in the region. In the present work, existing categories of handover, with a particular emphasis on those applicable to third-generation networks (3G), are studied and analyzed. One of the major classifications of handovers is that in hard and soft or softer handovers. An important advantage of soft handover is that it ensures the channel diversity which ultimately leads to handover algorithms that consume less energy. On the other hand, in the case of soft handover more network resources are required and greater complexity of the algorithms makes implementation harder. In this study, we studied the techniques used to decide if a handover should be executed in a mobile communication system. There are three basic categories of handover decision algorithms, i.e. those that are based on the level of the received signal power, the algorithms that are based on fuzzy logic and finally the priority algorithms. In the present study, we analyze critically the relevant literature on handover algorithms, emphasizing on those that can be applied in CDMA (Code Division Multiple Access) systems. In addition, we implement and simulate a model which combines "Grey Prediction" techniques with classical handover mechanisms to optimize the number of handovers performed while minimizing the call blocking probability. The simultaneous use and exploitation of three tools (Matlab, Simulink and Stateflow) simplifies the implementation and at the same time facilitates evaluation of the results. The key performance indicator used to evaluate algorithm’s performance is the number of handovers performed. Performed simulations are based on values of the parameters proposed in (Saeed Changiz Rezaei, Hossein Khalaj, 2005), combined with various values of Hysteresis. It is evident when looking at the simulation results, that the number of handovers performed decreases, as hysteresis increases. Finally, it should be noted that the very high sampling frequency of the model, combined with the smoothness of the mobile device motion and the good predictive ability of the Grey Prediction Algorithm, leads to the calculation of a very smooth sequence of consecutive values of the received signals and thus to the minimization of the number of handovers performed.

[pt] Com o crescimento gigantesco da Internet e dos sistemas de comunicações móveis celulares, as aplicações de processamento de voz nessas redes têm despertado grande interesse . Um problema particularmente importante nessa área consiste no reconhecimento de voz em um sistema servidor, baseado nos parâmetros acústicos calculados e quantizados no terminal do usuário (Reconhecimento de Voz Distribuído). Como em geral estes parâmetros não são os mais indicados como atributos de voz para o sistema de reconhecimento remoto, é importante que sejam examinadas diferentes transformações dos parâmetros, que permitam um melhor desempenho do reconhecedor. Esta dissertação trata da extração de atributos de reconhecimento eficientes a partir dos parâmetros dos codificadores utilizados em redes móveis celulares e em redes IP. Além disso, como a taxa dos parâmetros fornecidos ao reconhecedor de voz é normalmente superior àquela com a qual os codificadores geram os parâmetros, é importante analisar o efeito da interpolação dos parâmetros sobre o desempenho do sistema de reconhecimento, bem como o melhor domínio sobre o qual esta interpolação deve ser realizada. Estes são outros tópicos apresentados nesta dissertação. / [en] The huge growth of the Internet and cellular mobile communication systems has stimulated a great interest in the applications of speech processing in these networks. An important problem in this field consists in speech recognition in a server system, based on the acoustic parameters calculated and quantized in the user terminal (Distributed Speech Recognition). Since these parameters are not the most indicated ones for the remote recognition system, it is important to examine different transformations of these parameters, in order to allow a better performance of the recogniser. This dissertation is concerned with the extraction of efficient recognition features from the coder parameters used in cellular mobile networks and IP networks. In addition, as the rate that parameters supplied for the speech recogniser must be usually higher than that generated by the codec, it is important to analyze the effect of the interpolation of the parameters over the performance of the recognition system. Moreover, it is paramount to establish the best domain over which this interpolation must be carried out. These are other topics presented in this dissertation.