SIDS deaths exhibit a seasonal pattern with a winter peak, and the cause of this seasonality is unknown. The seasonal pattern is not symmetrical and it has been thought that the relatively flat winter peak may be due to the existence of more than one underlying population, where each population corresponds to a different cause of seasonality. In this thesis, mixtures of von Mises distributions have been fitted using maximum likelihood estimation to determine whether there is heterogeneity in the UK SIDS data. Various computational problems arise with the fitting procedures and attempts to tackle these for the SIDS data are discussed. A bootstrap likelihood ratio method is used to assess the number of components in the mixture, and its properties are investigated by simulation. Changes in the seasonal pattern since the 'back to sleep' campaign are also examined as any differences might give clues as to what caused the fall in 1992, and what the reasons for the remaining deaths might be. The von Mises distributions are compared with cosinor analysis and skewed regression models to determine the most appropriate method for modelling the seasonality in the data. Mixtures of Weibull and Gamma distributions are used to model the age distribution in SIDS. The motivation for this was to determine whether there are two or more groups of babies whose age-at-death distributions are different and to examine any changes since the 'back to sleep' campaign. Generalised linear models have previously been used to determine whether month of birth is an independent risk factor in addition to month of death and age at death. In this thesis, mixtures of these generalised linear models have been fitted using the EM algorithm to determine whether there are different groups of babies with different risks. Childhood type 1 diabetes mellitus is another condition which exhibits a seasonal pattern in diagnosis. The thesis concludes by considering analysis of these data using the mixture modelling approach.

This thesis investigates the effectiveness (doing the right thing) and efficiency (doing the thing right) of computer-based discrete-event simulation for designing manufacturing systems. This investigation looked at the use of this technology in the manufacture of discrete components in aerospace, automotive and consumer electrical (white goods) industries and for material handling (warehousing). Continuous and quasi-continuous manufacture have not been investigated and hence, the conclusions of this thesis cannot be generalised to cover these areas. Working hypotheses were developed and tested which looked at how discrete-event simulation influences the understanding of, confidence in and credibility of a system's design. Testing these working hypotheses lead to conclusions about how discrete-event simulation affects the quality of decision making and the lead-time to develop, commission and ramp-up a manufacturing facility. The following five factors were identified as influencing the efficiency of delivery of discrete-event simulation: l. Management of the simulation study and its intended benefits. 2. Management of customers' expectations. 3. Use of geometric animation. 4. Validation and establishing credibility. 5. How simulation's effectiveness varies over the life of a manufacturing system development project. A qualitative research methodology was employed to test these working hypotheses and to explore these efficiency factors. Twenty-three research subjects, in twelve companies, were drawn from the following three groups: ° Modellers (who provide the modelling service). ° Team members (who are closely involved with supporting the execution of a simulation study). Senior decision makers (who are not closely involved with the execution of the study, but who review its findings). A good practice model was developed for the efficient acquisition and application of the technology. This model consists of the following six elements: 1. Establishing and maintaining a DES modelling capability in the organisation. 2. Knowing whether to use DES modelling for this manufacturing system and when. 3. Defining the study's objectives and their means of measurement. 4. Specify responsibilities for supporting the study and implementing its findings. 5. There is no improvement in the effectiveness of DES modelling in using 2D rather than 3D geometric animation. 6. 3D geometric animation can increase the efficiency of a study, if used appropriately. Conclusions were made about the effectiveness of discrete-event simulation, how the above mentioned factors influence its efficiency of delivery and how to implement the good practice model.

Abstract Grinding mills are generally very inefficient, difficult to control and costly, in terms of both power and steel consumption. Improved understanding of temperature behaviour in milling circuits can be used in the model-based control of milling circuits. The loss of energy to the environment from the grinding mill is significant hence the need for adequate modeling. The main objectives of this work are to quantify the various rates of energy loss from the grinding mill so that a reliable model for temperature behaviour in a mill could be developed. Firstly models of temperature behaviour in a grinding mill are developed followed by the development of a model for the overall heat transfer coefficient for the grinding mill as a function of the load volume, mill speed and the design of the liners and mill shell using the energy balances in order to model energy loss from the mill. The energy loss via convection through the mill shell is accounted for by quantifying the overall heat transfer coefficient of the shell. Batch tests with balls only were conducted. The practical aspect of the work involved the measurement of the temperatures of the mill load, air above the load, the liners, mill shell and the environmental temperature. Other measurements were: mill power and sound energy from the mill. Energy balances are performed around the entire mill. A model that can predict the overall heat transfer coefficient over a broad range of operating conditions was obtained. It was found that the overall heat transfer coefficient for the grinding mill is a function of the individual heat transfer coefficients inside the mill and outside the mill shell as well as the design of the liners and shell. It was also found that inside heat transfer coefficients are affected by the load volume and mill speed. The external heat transfer coefficient is affected by the speed of the mill. The values for the overall heat transfer coefficient obtained in this work ranged from 14.4 – 21W/m2K. iv List of Publications The author has published the following papers based on the contents of this dissertation as follows: Published conference abstract Kapakyulu, E., and Moys, M.H., 2005. Modelling of energy loss to the environment from the grinding mill, Proceedings of the Mineral Processing 2005’ Conference, SAIMM, Cape Town, South Africa, 4-5 Aug. pp 65-66 - SP03 Research Papers: Accepted for publication and currently in press in Minerals Engineering: Kapakyulu, E., and Moys, M.H., 2006. Modelling of energy loss to the environment from a grinding mill, Part I: Motivation, Literature Survey and Pilot Plant Measurements, (Currently in press in Minerals Engineering) Kapakyulu, E., and Moys, M.H., 2006. Modelling of energy loss to the environment from a grinding mill, Part II: Modeling the overall heat transfer coefficient, (Currently in press in Minerals Engineering)

The seven transmembrane α-helices of the human 5 -HT_1A , 5-HT_1Dα and 5-HT_1Dβ receptors have been modelled using the 3-dimensional coordinates of the seven transmembrane a-helices of the bacterial protein bacteriorhodopsin as a template. The probable 5-HT binding site was identified between helices 3, 4, 5 and 6. Interactions between the natural ligand 5-HT (A) and the receptor models are described in detail and the agonist binding site is further validated by the docking of four known 5-HT receptor ligands. The models are able to account qualitatively for the receptor binding affinities of the studied ligands. Small molecule similarity studies suggest that a thieno[2,3-b]pyridine analog (B) of 5-HT could possibly act as a bio-isostere for 5-HT. This was further corroborated when (B) was docked into the 5-HT receptor models and was found to be accommodated easily in the 5-HT binding site participating in the same interactions as observed for 5-HT. Thieno[2,3-b]pyridines similar to (B) were thus identified as synthetic target compounds. Furthermore, the models were used to provide suggestions for the design of novel, more selective, 5-HT receptor agonists. The thieno[2,3-b]pyridine ester (C; R=C0₂Et) was reduced to the hydroxymethyl derivative (C; R=CH₂0H) but the methylthio group could not be successfully removed in the presence of the thiophene ring. Using a different approach the thieno[2,3-b]pyridine (D; R=CO₂Me, X=CN) was synthesised as a model compound and converted through to the <i>t</i>-BOC protected amines (D, R=NHCO₂C(CH₃)₃, X=CN) and (D, R=NHCO₂C(CH₃)₃ , X=CONH₂). The same reactions were applied to ethyl-3-(5-cyanothieno[2,3-b] pyridin-3-yl)propanoate (E, R=CO₂Et, X=CN) but this unfortunately could not be converted through to the required <i>t</i>-BOC protected 5-HT analog (E, R=NHCO₂C(CH₃)₃, X=CN).

The motivation for this work arose from the recognised inability of relational databases to store and manipulate data that is outside normal commercial applications (e.g. graphical data). The published work in this area is described with respect to the major problems of representation and manipulation of complex data. A general purpose data model, called GDB, that sucessfully tackles these major problems is developed from a formal specification in ML and is implemented using the PRECI/C database system. This model uses three basic graphical primitives (line segments, plane surfaces - facets, and volume elements tetrons) to construct graphical objects and it is shown how user designed primitives can be included. It is argued that graphical database query languages should be designed to be application specific and the user should be protected from the relational algebra which is the basis of the database operations. Such a base language (an extended version of DEAL) is presented which is capable of performing the necessary graphical manipulation by the use of recursive functions and views. The need for object hierarchies is established and the power of the DEAL language is shown to be necessary to handle such complex structures. The importance of integrity constraints is discussed and some ideas for the provision of user defined constraints are put forward.

La maladie de Huntington (MH) est une maladie neurodégénérative héréditaire. Elle est causée par l'expansion d'un motif CAG qui codant une séquence de poly-glutamine (polyQ) dans le gène huntingtine (HTT). La survenue des symptômes de la maladie est tardive. Pour étudier la MH, un grand nombre de modèles génétiques animaux et cellulaires ont été utilisés au cours des 15 dernières années sans se traduire en traitements bénéfiques de façon durable pour les patients. Les médicaments disponibles servent actuellement à gérer les symptômes de la maladie, et n’ont pas d’effet sur le pronostic létal. Les modèles existants semblent incapables de répliquer pleinement les premières anomalies transcriptionnelles intervenant dans la MH. J’ai tiré profit de la disponibilité de lignées de cellules hESC-MH et émis l'hypothèse qu’elles sont adaptées au déchiffrement des mécanismes pathologiques ou correcteurs impliqués dans la phase pré-symptomatique de la MH, du développement embryonnaire à la vie adulte. J'ai développé des protocoles et des outils cellulaires et moléculaires pour établir les hESC mutantes et normales comme modèle cellulaire de la MH. Puis j’ai utilisé ces cultures pour explorer par une approche transcriptomique la dérégulation transcriptionnelle dans les cellules neurales immatures porteuses de la mutation causale de la MH. J’ai identifié de nouveaux biomarqueurs cellulaires précoces associés à la mutation causale de la MH. Ces résultats suggèrent l’existence de mécanismes moléculaires encore inconnus et spécifiques des phases les plus précoces de la maladie. Ces découvertes pourraient fournir des cibles originales pour une intervention pharmacologique pré-symptomatique. / Huntington’s Disease (HD) is a neurodegenerative inherited disease. It is caused by the extension of a CAG motif coding for a poly-glutamine (polyQ) tract within huntingtin gene (HTT). Symptoms declaration occurs late in life. In order to better understand HD a wealth of cellular and animal genetic models both have fuelled 15 years of studies, but never transduced into truely lasting and beneficial pharmacological treatment. Avalaible drugs are used to handle symptomatology, but don’t address the lethal prognosis. Existing models seems to fail at replicating fully the first steps of transcriptionnal alterations occurring in HD. I took advantage of the recent disponibility of hESC naturally carrying HD mutation to investigate the hypothesis that they can yield insight into early presymptomatic mechanisms of HD from embryonic development to adulthood. I developed protocols, cellular and molecular tools to establish HD and normal hESC as HD cellular models. Then I used these tools to investigate transcriptomic dysregulation in immature neural cells carrying the HD causing mutation. I identified new early cellular biomarkers associated to the mutation. These results suggest the existence of previously unknown molecular mécanisms spécific to the earliest stages of the disease. These new biomarkers could be very interesting targets for pre-symptomatic pharmacological intervention.

Light signalling mechanisms are crucial for plant adaptation to environmental changes: it has been found that both light and temperature cues are interpreted by closely interacting signalling pathways. In Arabidopsis thaliana, the seedling hypocotyl has emerged as an ideal model system to study light and temperature control of cell expansion. The light sensitivity is affected by the light intensity (fluence rate), an illustration of which is exemplified in the fluence rate response curve, and also the quality of that light (i.e. the ratio of red: far-red wavelengths of light). Here I have elucidated the phytochrome B (phyB) signalling mechanism in great detail and subsequently created a mathematical model of the red-light signalling pathway. This model is further extended to include changes within the ambient temperature range. Through application of stochastic modelling techniques it was possible to demonstrate the importance of dark reversion within the phyB mechanism in a diurnal cycle. Further, by considering alternative modelling methods and applying a highly integrated mathematical and experimental approach, it has been possible to show how the integration of light and temperature signals drives a molecular switch in transcriptional activity and hence control of cell expansion. This work highlights the importance of a multidisciplinary approach to molecular biology and describes the significance of signal convergence on molecular signalling.

L’intérêt pour la modélisation et la commande des engins volants s’est accru de manière significative au cours de ces dernières années. La complexité et les possibilités des engins volants s’accroissent rapidement et la gamme des missions qu’elles doivent réaliser se développe. Cependant afin que les drones puissent atteindre ce potentiel, certains défis techniques doivent être surmontés, notamment l’étude et l’intégration de la flexibilité structurelle, la prise en compte des phénomènes aérodynamiques, et l’élaboration de stratégies de commandes adaptées. Le travail présenté s’inscrit dans ce cadre et porte spécifiquement sur deux types de drones : -Les plus légers que l’air : Application à la modélisation et commande d’un dirigeable souple. -Les plus lourds que l’air : Application à la modélisation et commande d’un quadrirotor flexible : le XSF. Nous présentons dans un premier temps un modèle global d’engins volants flexibles autonomes. On admet que ces objets volants subissent de grands déplacements et de petites déformations élastiques. Le formalisme utilisé est basé sur l’approche de Newton-Euler, approche souvent utilisée dans le cas d’objets volants rigides. Dans cette étude nous généralisons le formalisme de corps rigides existant en y incluant l’effet de la flexibilité, sans pour autant détruire la méthodologie globale, et ce au moyen d’une technique hybride lagrangienne-eulerienne. La flexibilité apparaît dans le système dynamique global par le moyen d’un nombre réduit de degrés de liberté supplémentaires issus d’une synthèse modale. Cette technique permet de faciliter par la suite l’élaboration d’algorithmes de commande et de stabilisation. Le phénomène des masses ajoutées est également pris en considération. Une méthode originale de traitement analytique de ce phénomène a été établie pour un corps flexible en grands mouvements. Elle est basée sur la notion de potentiel flexible, et sur le développement de l’énergie cinétique du fluide sous l’effet d’un mouvement global du corps flexible. Cette méthode a permis de mettre en évidence le couplage rigide-flexible dans la matrice des masses ajoutées pour un traitement global d’un dirigeable flexible. On présente aussi le modèle dynamique et aérodynamique du quadrirotor flexible XSF conçu au laboratoire IBISC et destiné à un concours interuniversitaire sur les microdrones. Une technique robuste " Backstepping " est réalisée pour la stabilisation du dirigeable flexible au voisinage d’un point cible. Et une stratégie de contrôle de PID a été proposée pour la stabilisation de l’XSF. La stratégie de commande est contrainte par l’impératif d’optimisation du rapport précision/portabilité, pour que les algorithmes développés puissent être intégrés dans l’informatique embarquée de ces engins volants. Une validation numérique est présentée à la fin du rapport. / The interest of dynamic modeling and control of the autonomous flying objects increased significantly during these last years. Complexity and capability of these flying objects are expanding rapidly now, and the range of missions their designed to support is growing. In order to fulfil this requirement, it is necessary, in one hand, to introduce the effect of the structural flexibility and the aerodynamic phenomenon in the dynamic model, and in the other hand, to build a suitable strategy of command and stabilization for these flying objects. The work is registered within this framework, and relates two types of engines: - Lighter than air vehicle: Application on the modelling and control of flexible airship. - Heavier than air vehicle: Application on the modelling and control of flexible micro-drone. We present a general model of autonomous flexible flying engine undergoing great overall motion and small elastic displacements. The formalism used is based on the Newton- Euler approach, which is frequently used for rigid flying objects. In this study we generalize the existing formalisms for rigid bodies, by including the effect of flexibility without destroying the total methodology. A modal synthesis is used. A hybrid method based on the energetic principles and Lagrange equations is presented. The phenomenon of the added masses is also taken into account. In order to integrate the fluid-structure interaction of a flexible airship, we develop an original analytical formulation of the problem using both the new notion of flexible flow potential, and the development of the kinetic energy of the air constrained by the motion and the vibration of the airship. This method allows to put in an obvious place the coupling "overall motion - flexibility" in the added mass matrix. We also present the dynamic and aerodynamic model of a flexible quadrirotor called XSF, designed in the IBISC laboratory and intended for an interuniversity competition for microdrones. A robust methodology based on the " Backstepping " control is realized to stabilize the airship around a desired position, and a PID controller is proposed to stabilise the XSF. The strategy of command is compelled by imperative to optimize the ratio precision/portability, to allow an easy insertion of the developed algorithms in the embedded electronics. A numerical validation is presented in the end of this report.

This thesis documents the development and application of a computer model for gibbsite, an aluminium tri-hydroxide polymorph. In particular, the work has emphasized the idea of computer modelling techniques combining with ex- observations to provide greater insight than either could separately. Chapter One provides an overview and introduction to the fields of solid state chemistry, crystallization and computer modelling. These ideas are extended in Chapter Two to include a more detailed discussion of the theoretical principles behind the modelling in this project. The development of transferable oxalate and hydroxide potential models, intended primarily for sodium oxalate and gibbsite, is described in Chapter Three. Both ab initio hypersurface fitting and lattice fitting techniques were utilized, with an average structural fitting error of under two percent. In addition, the potentials were used to successfully reproduce several (related) crystal structures, thus establishing the quality of the model. In Chapter Four, the model for gibbsite was employed in generating equilibrium and growth morphologies. The equilibrium morphology was found to give excellent agreement with experiment, with all observed faces present. However, the importance of the prismatic planes is underestimated. Also discussed in the chapter is a method for predicting the phenomenon of crystalline twinning. This technique was successfully applied to a number of systems, including gibbsite and sodium oxalate. In Chapter Five, the equilibrium morphology calculations performed earlier were extended by probing the effects of cation incorporation on the habit of gibbsite. This study was conducted in order to provide a first step in estimating the role of the crystallizing solution. Calculations of the change in surface energy caused by the replacement of a surface proton with a cation from solution ++ / were made. Different crystal habits were constructed by applying a range of defect surface coverage values to each of the faces appearing in the morphology. The resulting defect morphologies were in excellent agreement with crystal habits commonly observed by experimentalists. Also, the work provided an explanation for the earlier underestimation of the prismatic faces. Chapter Six documents molecular simulations of solutions containing the major species known to be present in industrial and experimental Bayer liquors. The structuring in two solutions, one containing sodium hydroxide and the other potassium hydroxide, was probed by constructing graphs of the radial distribution functions. These plots indicated that a significant degree of ion pairing was occurring between the alkali metal cations (Na+ and K+) and the aluminate monomer ([Al(OH)4(subscript)]-). Furthermore, these cations were found to be acting as 'bridges' which stabilize multiple aluminate monomers, allowing them to form clusters. This data was used to assist in explaining vibrational spectra, and to postulate that clustering may be the origin of the fine particle suspensions noted during the induction period.

After the Boxing Day 2004 Sumatran Tsunami, a review of tsunami hazard and risk for New Zealand identified Gisborne as the urban area with the greatest risk. Gisborne could experience gt;500 fatalities and extensive damage to infrastructure during a severe tsunami. The severity of a tsunami is likely to be low for distance sources given the effectiveness of the Pacific Tsunami Warning System. However, there is a substantial risk from local sources, as no local warning system of any kind exists. Prompt evacuation is probably the most cost-effective tsunami mitigation strategy available for New Zealand coastal locations, including Gisborne. This requires both knowledge of the extent of tsunami inundation, and sufficient warning of the tsunami arrival. Hence, there are two main objectives for this investigation: 1. Determine the likely extent of tsunami inundation for Gisborne City and surrounding populated coastal locations in Poverty Bay, using a combination of hydrodynamic tsunami modelling and GIS. The modelling will simulate historical events, particularly the largest historical tsunami, the May 1947 local tsunami. Modelling will consider potential events based on the Maximum Credible Earthquake for local sources associated with the Hikurangi Deformation Front. 2. Create inundation maps of Poverty Bay that can be used for future town planning and emergency plans.