In this thesis, the Erlang queueing model Af/i/l, where customers arrive at random mean rate A and service times have an Erlang distribution with parameter k and iro service rate u, has been considered from different perspectives. Firstly, an analytic metl of obtaining the time-dependent probabilities, pn,,(() for the M/Ek/l system have t> proposed in terms of a new generalisation of the modified Bessel function when initk there are no customers in the system. Results have been also generalised to the case wl initially there are a customers in the system. Secondly, a new generalisation of the modified Bessei function and its generating function have been presented with its main properties and relations to other special functii (generalised Wright function and Mittag-Leffler function) haw been noted. Thirdly, the mean waiting tune in the queue, H',(f), has been evaluated, using Lucha results. The double-exponential approximation of computing Yq(t) has been proposed different values of p. which gives results within about % of the 'exact1 values obtained fr numerical solution of the differential-difference equations. The advantage of this approximation is that it provides additional information, via its functional form of the characterisl of the transient solution. Fourthly, the inversion of the Laplace transform with the application to the queues 1 been studied and verified for A//A//1 and M/Ek/l models of computing Wq{t}. Finally, an application of the A//fi/l queue has been provided in the example of hour traffic flow for the Severn Bridge. One of the main reasons for studying queue models from a theoretical point of view is to develop ways of modelling real-life system. The analytic results have been confirmed with the simulation.

The Critical Care Unit (CCU) is the sector of the hospital where, as the name suggests, critically ill patients receive treatment. The main aim of this research is to identify and apply suitable Operational Research techniques to model patient flow in the CCU at the University Hospital of Wales, Cardiff. The Operational Research techniques employed in this thesis include queueing theory and simulation. These methods have been utilised previously in the field of healthcare with much success. The thesis begins by considering two aspects of queueing theory, namely batch service queueing theory and batch arrival queueing theory. The latter of these is utilised to model patient flow within the CCU. Although queueing theory may be used as a good approximation to activities in the Unit, it does not incorporate all aspects of real-life. Thus discrete-event simulation is suggested as an alternative approach. Two types of statistical analysis, CART and Regression, are applied to both length of stay and mortality variables. The results from these statistical tests are compiled and investigated in more depth. Finally, a discrete event simulation model is built in Visual Basic for Applications, for Microsoft Excel. This simulation model incorporates many of the complexities of a CCU, such as patient priority and cancellation of scheduled patients if all beds on the Unit are occupied. The model is then used to test various "what-if type" scenarios, including the possibility of funding additional beds, the concept of ring-fencing of beds for different levels of care, and the likely effect of reducing the impact of bed-blocking.

Critical Care Units (CCUs) are one of the most complex and expensive of all medical resources and hospital managers are challenged to meet the demand for critical care services with adequate capacity. The pressure on critical care beds is continuously increasing as new medical equipment provides the opportunity to save more patients lives. It is therefore crucial that beds are managed well and used efficiently. This thesis describes two major projects, the first undertaken in conjunction with the CCU at the University Hospital of Wales in Cardiff (UHW); and the second with two CCUs from the Aneurin Bevan Health Board. In the first project data has been analysed to determine the flow of patients through the Unit. Admissions to CCUs were categorised under two headings: emergency, and elective. The length of stay in the CCU is heavily dependent on the admission category. In this thesis, both computer simulation and theoretical queueing models have been considered, which show how improvements in bed management may be achieved by considering these two categories of patients separately. The vast majority of previous literature in this field is concerned only with steady-state conditions, whereas in reality the processes are time-dependent. This thesis goes some way to addressing this deficiency. The second project relates to work undertaken with managers from the Royal Gwent Hospital in Newport and at the Nevill Hall Hospital in Abergavenny. Data from both hospitals have been analysed to define arrival and service processes. A state-dependent theoretical queueing model has been considered which has been used to investigate the significance of combining the two units. The model has been also utilised to advise on the number of beds the new combined unit should have in order to satisfy targets quoted by the hospital managers. In the final part of the thesis, consideration has been given to the impact of collaboration, or lack thereof, between hospitals using a game theoretical approach. The effect of patient diversion has been studied. To formally investigate the impact of patients transfers, a Markov chain model of the two CCUs has been set-up, each admitting two arrival streams: namely, their own patients and transfers from other hospital. Four different models were considered and for each model the effect of targets, demand and capacity were studied. The efficiency of a system which degrades due to selfish behaviour of its agents has been measured in terms of Price of Anarchy.

This thesis discusses queueing systems in which decisions are made when customers arrive, either by individual customers themselves or by a central controller. Decisions are made concerning whether or not customers should be admitted to the system (admission control) and, if they are to be admitted, where they should go to receive service (routing control). An important objective is to compare the effects of "selfish" decision-making, in which customers make decisions aimed solely at optimising their own outcomes, with those of "socially optimal" control policies, which optimise the economic performance of the system as a whole. The problems considered are intended to be quite general in nature, and the resulting findings are therefore broad in scope. Initially, M/M/1 queueing systems are considered, and the results presented establish novel connections between two distinct areas of the literature. Subsequently, a more complicated problem is considered, involving routing control in a system which consists of heterogeneous, multiple-server facilities arranged in parallel. It is shown that the multiple-facility system can be formulated mathematically as a Markov Decision Process (MDP), and this enables a fundamental relationship to be proved between individually optimal and socially optimal policies which is of great theoretical and practical importance. Structural properties of socially optimal policies are analysed rigorously, and it is found that 'simple' characterisations of socially optimal policies are usually unattainable in systems with heterogeneous facilities. Finally, the feasibility of finding 'near-optimal' policies for large scale systems by using heuristics and simulation-based methods is considered.

As the prominence of the service sector is increasing in developed nations, new and exciting opportunities are arising for operational researchers to develop and apply models which offer managers solutions to improve the quality of their services. The development of time-dependent stochastic models to analyse complex service systems and generate effective personnel schedules are key to this process, enabling organisations to strike a balance between the provision of a good quality service whilst avoiding unnecessary personnel costs. Specifically within the healthcare sector, there is a need to promote efficient management of an Emergency Medical Service (EMS), where the probability of survival is directly related to the speed of assistance. Motivated by case studies investigating the operation of the Welsh Ambulance Service Trust (WAST), this thesis aims to investigate how operational research (OR) techniques can be developed to analyse priority service systems subject to demand that is of an urgent nature, cannot be backlogged, is heavily time-dependent and highly variable. A workforce capacity planning tool is ultimately developed that integrates a combination of forecasting, queueing theory, stochastic modelling and optimisation techniques into a single spreadsheet model in order to predict future demand upon WAST, set staffing levels, and optimise shift schedules and rosters. The unique linking together of the techniques in a planning tool which further captures time-dependency and two priority classes enables this research to outperform previous approaches, which have generally only considered a single class of customer, or generated staffing recommendations using approximation methods that are only reliable under limited conditions. The research presented in this thesis is novel in several ways. Primarily, the first section considers the potential of a nonparametric modelling technique known as Singular Spectrum Analysis (SSA) to improve the accuracy of demand forecasts. Secondly, the main body of work is dedicated to adapting numerical queueing theory techniques to accurately model the behaviour of time-dependent multi-server priority systems across shift boundaries and evaluate the likelihood of excessive waits for service for two customer classes. The final section addresses how shifts can be optimally scheduled using heuristic search techniques. The main conclusions are that in addition to offering a more flexible approach, the forecasts generated by SSA compare favourably to those obtained using traditional methods, and both approximate and numerical modelling techniques may be duly extended to set staffing levels in complex priority systems.

The practicality of the stochastic network calculus (SNC) is often questioned on grounds of looseness of its performance bounds. The reason for its inaccuracy lies in the usage of too elementary tools from probability theory, such as Boole’s inequality, which is unable to account for correlations and thus inappropriate to properly model arrival flows. In this thesis, we propose an extension of stochastic network calculus that characterizes its main objects, namely arrival and service processes, in terms of martingales. This characterization allows to overcome the shortcomings of the classical SNC by leveraging Doob’s inequality to provide more accurate performance bounds. Additionally, the emerging stochastic network calculus with martingales is quite versatile in the sense that queueing related operations like multiplexing and scheduling directly translate into operations of the corresponding martingales. Concretely, the framework is applied to analyze the per-flow delay of various scheduling policies, the performance of random access protocols, and queueing scenarios with a random number of parallel flows. Moreover, we show our methodology is not only relevant within SNC but can be useful also in related queueing systems. E.g., in the context of multi-server systems, we provide a martingale-based analysis of fork-join queueing systems and systems with replications. Throughout, numerical comparisons against simulations show that the Martingale bounds obtained with Doob’s inequality are not only remarkably accurate, but they also improve the Standard SNC bounds by several orders of magnitude.

L’objectif de cette thèse est d’étudier, sous divers aspects, le processus de formation d’amyloïde à partir de la polymérisation de protéines. Ces phénomènes, aussi bien in vitro que in vivo, posent des questions de modélisation mathématique. Il s’agit ensuite de conduire une analyse des modèles obtenus. Dans la première partie nous présentons des travaux effectués en collaboration avec une équipe de biologistes. Deux modèles sont introduits, basés sur la théorie en vigueur du phénomène Prions, que nous ajustons aux conditions expérimentales. Ces modèles nous permettent d’analyser les données obtenues à partir d’expériences conduites en laboratoire. Cependant celles-ci soulèvent certains phénomènes encore inexpliqués par la théorie actuelle. Nous proposons donc un autre modèle qui corrobore les données et donne une nouvelle approche de la formation d’amyloïde dans le cas du Prion. Nous terminons cette partie par l’analyse mathématique de ce système compose d’une infinité d’équations différentielles. Ce dernier consiste en un couplage entre un système de type Becker-Doring et un système de polymérisation-fragmentation discrète. La seconde partie s’attache à l’analyse d’un nouveau modèle pour la polymérisation de protéines dont la fragmentation est sujette aux variations du fluide environnant. L’idée est de décrire au plus près les conditions expérimentales mais aussi d’introduire de nouvelles quantités macroscopiques mesurables pour l’étude de la polymérisation. Le premier chapitre de cette partie présente une description stochastique du problème. On y établit les équations du mouvement des polymères et des monomères (de type Langevin) ainsi que le formalisme pour l’étude du problème limite en grand nombre. Le deuxième chapitre pose le cadre fonctionnel et l’existence de solutions pour l équation de Fokker-Planck- Smoluchowski décrivant la densité de configuration des polymères, elle-même couplée a une équation de diffusion pour les monomères. Le dernier chapitre propose une méthode numérique pour traiter ce problème. On s’intéresse dans la dernière partie à la modélisation de la maladie d’Alzheimer. On construit un modèle qui décrit d’une part la formation de plaque amyloïde in vivo, et d’autre part les interactions entre les oligomères d’Aβet la protéine prion qui induiraient la perte de mémoire. On mène l’analyse mathématique de ce modèle dans un cas particulier puis dans un cas plus général ou le taux de polymérisation est une loi de puissance / The aim of this thesis is to study, under several aspects, the formation of amyloids from proteins polymerization. The mathematical modelling of these phenomena in the case of in vitro or in vivo polymerisation remains questioned. We then propose here several models, which are also investigated from theoritical and numerical point of view. In the first part we present works done in collaboration with biologists. We propose two models based on the current theory on Prion phenomena that are designed for specific experimental conditions. These models allow us to analyse the experimental data obtained in laboratory and raise phenomena that remain unexplained by the theory. Then, from these results and biophysical considerations, we introduce a model which corroborates with data and provides a new approach on the amyloid formation in the particular case of Prion. This part is ended by the mathematical analysis of the model consisting of an infinite set of differentials equations. The system analysed is a Becker-Doring system coupled to a discrete growth-fragmentation system. The second part is dedicated to the analysis of a new model for polymerization of proteins with fragmentation subject to the surrounding variations of the fluid. Thus, we propose a model which is close to the experimental conditions and introduce new measurable macroscopic quantities to study the polymerization. The first introductory chapter states the stochastic description of the problem. We give the equations of motion for each polymers and monomers as well as a general formalism to study the limit in large number. Next, we give the mathematical framework and prove the existence of solutions to the Fokker-Planck-Smoluchowski equation for the configurational density of polymers coupled to the diffusion equation for monomers. The last chapter provides a numerical method adapted to this problem with numerical simulations In the last part, we are interested in modelling Alzheimer’s disease. We introduce a model that describes the formation of amyloids plaques in the brain and the interactions between Aβ-oligomers and Prion proteins which might be responsible of the memory impairment. We carry out the mathematical analysis of the model. Namely, for a constant polymerization rate, we provide existence and uniqueness together with stability of the equilibrium. Finally we study the existence in a more general and biological relevant case, that is when the polymerization depends on the size of the amyloid

The purpose of this research is to investigate and analyse some batch arrival queueing systems with Bernoulli scheduled vacation process and single server providing service. The study aims to explore and extend the work done on vacation and unreliable queues with a combination of assumptions like balking and re-service, reneging during vacations, time homogeneous random breakdowns and fluctuating modes of service. We study the steady state properties, and also transient behaviour of such queueing systems. Due to vacations the arriving units already in the system may abandon the system without receiving any service (reneging). Customers may decide not to join the queue when the server is in either working or vacation state (balking). We study this phenomenon in the framework of two models; a single server with two types of parallel services and two stages of service. The model is further extended with re-service offered instantaneously. Units which join the queue but leave without service upon the absence of the server; especially due to vacation is quite a natural phenomenon. We study this reneging behaviour in a queueing process with a single server in the context of Markovian and non-Markovian service time distribution. Arrivals are in batches while each customer can take the decision to renege independently. The non-Markovian model is further extended considering service time to follow a Gamma distribution and arrivals are due to Geometric distribution. The closed-form solutions are derived in all the cases. Among other causes of service interruptions, one prime cause is breakdowns. We consider breakdowns to occur both in idle and working state of the server. In this queueing system the transient and steady state analysis are both investigated. Applying the supplementary variable technique, we obtain the probability generating function of queue size at random epoch for the different states of the system and also derive some performance measures like probability of server‟s idle time, utilization factor, mean queue length and mean waiting time. The effect of the parameters on some of the main performance measures is illustrated by numerical examples to validate the analytical results obtained in the study. The Mathematica 10 software has been used to provide the numerical results and presentation of the effects of some performance measures through plots and graphs.

La question du rachat préoccupe les assureurs depuis longtemps notamment dans le contexte des contrats d'épargne en Assurance-Vie, pour lesquels des sommes colossales sont en jeu. L'émergence de la directive européenne Solvabilité II, qui préconise le développement de modèles internes (dont un module entier est dédié à la gestion des risques de comportement de rachat), vient renforcer la nécessité d'approfondir la connaissance et la compréhension de ce risque. C'est à ce titre que nous abordons dans cette thèse les problématiques de segmentation et de modélisation des rachats, avec pour objectif de mieux connaître et prendre en compte l'ensemble des facteurs-clefs qui jouent sur les décisions des assurés. L'hétérogénéité des comportements et leur corrélation ainsi que l'environnement auquel sont soumis les assurés sont autant de difficultés à traiter de manière spécifique afin d'effectuer des prévisions. Nous développons ainsi une méthodologie qui aboutit à des résultats très encourageants ; et qui a l'avantage d'être réplicable en l'adaptant aux spécificités de différentes lignes de produits. A travers cette modélisation, la sélection de modèle apparaît comme un point central. Nous le traitons en établissant les propriétés de convergence forte d'un nouvel estimateur, ainsi que la consistance d'un nouveau critère de sélection dans le cadre de mélanges de modèles linéaires généralisés / Insurers have been concerned about surrenders for a long time especially in Saving business, where huge sums are at stake. The emergence of the European directive Solvency II, which promotes the development of internal risk models (among which a complete unit is dedicated to surrender risk management), strengthens the necessity to deeply study and understand this risk. In this thesis we investigate the topics of segmenting and modeling surrenders in order to better know and take into account the main risk factors impacting policyholders’ decisions. We find that several complex aspects must be specifically dealt with to predict surrenders, in particular the heterogeneity of behaviours and their correlations as well as the context faced by the insured. Combining them, we develop a methodology that seems to provide good results on given business lines, and that moreover can be adapted for other products with little effort. However the model selection step suffers from a lack of parsimoniousness: we suggest to use another criteria based on a new estimator, and prove its consistant properties in the framework of mixtures of generalized linear models

Cette thèse est consacrée au développement de modèles mathématiques de l'hématopoïèse et de maladies du sang. Elle traite du développement de modèles hybrides discrets continus et de leurs applications à la production de cellules sanguines (l'hématopoïèse) et de maladies sanguines telles que le lymphome et le myélome. La première partie de ce travail est consacrée à la formation de cellules sanguines à partir des cellules souches de la moelle osseuse. Nous allons principalement étudier la production des globules rouges, les érythrocytes. Chez les mammifères, l'érythropoïèse se produit dans des structures particulières, les îlots érythroblastiques. Leur fonctionnement est régi par de complexes régulations intra et extracellulaire mettant en jeux différents types de cellules, d'hormones et de facteurs de croissance. Les résultats ainsi obtenus sont comparés avec des données expérimentales biologiques ou médicales chez l'humain et la souris. Le propos de la deuxième partie de cette thèse est de modéliser deux maladies du sang, le lymphome lymphoblastique à cellules T (T-LBL) et le myélome multiple (MM), ainsi que leur traitement. Le T-LBL se développe dans le thymus et affecte la production des cellules du système immunitaire. Dans le MM, les cellules malignes envahissent la moelle osseuse et détruisent les îlots érythroblastiques empêchant l'érythropoïèse. Nous développons des modèles multi-échelles de ces maladies prenant en compte la régulation intracellulaire, le niveau cellulaire et la régulation extracellulaire. La réponse au traitement dépend des caractéristiques propres à chaque patient. Plusieurs scénarios de traitements efficaces, de rechutes et une résistance au traitement sont considérés. La dernière partie porte sur un modèle d'équation de réaction diffusion qui peut être utilisé pour décrire l'évolution darwinienne des cellules cancéreuses. L'existence de “pulse solutions”, pouvant décrire localement les populations de cellules et leurs évolutions, est prouvée / The thesis is devoted to mathematical modeling of hematopoiesis and blood diseases. It is based on the development of hybrid discrete continuous models and to their applications to investigate production of blood cell (hematopoiesis) and blood diseases such as lymphoma and myeloma. The first part of the thesis concerns production of blood cells in the bone marrow. We will mainly study production of red blood cells, erythropoiesis. In mammals erythropoiesis occurs in special structures, erythroblastic islands. Their functioning is determined by complex intracellular and extracellular regulations which include various cell types, hormones and growth factors. The results of modeling are compared with biological and medical data for humans and mice. The purpose of the second part of the thesis is to model some blood diseases, T cell Lymphoblastic lymphoma (T-LBL) and multiple myeloma (MM) and their treatment. TLBL develops in the thymus and it affects the immune system. In MM malignant cells invade the bone marrow and destroy erythroblastic islands preventing normal functioning of erythropoiesis. We developed multi-scale models of these diseases in order to take into account intracellular molecular regulation, cellular level and extracellular regulation. The response to treatment depends on the individual characteristics of the patients. Various scenarios are considered including successful treatment, relapse and development of the resistance to treatment. The last part of the thesis is devoted to a reaction-diffusion model which can be used to describe Darwinian evolution of cancer cells. Existence of pulse solutions, which can describe localized cell populations and their evolution, is proved