Adrian C. Melissinos. / "November 10, 1958." Reprinted from The physical review, vol. 115, no. 1, pp. 126-129, July 1, 1959." / Includes bibliographical references. / Army Signal Corps Contract DA36-039-sc-78108. Dept. of the Army Task 3-99-20-001 and Project 3-99-00-000.

We develop an extension to the tactical planning model (TPM) for a job shop by the third author. The TPM is a discrete-time model in which all transitions occur at the start of each time period. The time period must be defined appropriately in order for the model to be meaningful. Each period must be short enough so that a job is unlikely to travel through more than one station in one period. At the same time, the time period needs to be long enough to justify the assumptions of continuous workflow and Markovian job movements. We build an extension to the TPM that overcomes this restriction of period sizing by permitting production control over shorter time intervals. We achieve this by deriving a continuous-time linear control rule for a single station. We then determine the first two moments of the production level and queue length for the workstation. / Singapore-MIT Alliance (SMA)

Thesis (Ph. D.)--Ohio State University, 2006. / Title from first page of PDF file. Includes bibliographical references (p. 129-135).

Ce travail a pour but la réalisation d'un outil permettant d'estimer la germination et la distribution en taille de cristaux d'un procédé de cristallisation. Pour ce faire un modèle de cristallisation a été développé et nous a permis de réaliser un observateur de dimension finie afin d'estimer la germination. Par la suite, un observateur de dimension infinie a été mis au point dans le but d'estimer la distribution de taille des cristaux à partir de la germination et de la mesure d'une taille de la distribution. Les deux observateurs ont été validés expérimentalement sur un procédé "batch" de cristallisation de l'oxalate d'ammonium.

In this thesis we study the relativistic multipole moments for stationary asymptotically flat spacetimes as introduced by Geroch and Hansen. These multipole moments give an asymptotic description of the gravitational field in a coordinate independent way. Due to this good description of the spacetimes, it is natural to try to construct a spacetime from only the set of multipole moments. Here we present a simple method to do this for the static axisymmetric case. We also give explicit solutions for the cases where the number of non-zero multipole moments are finite. In addition, for the general stationary axisymmetric case, we present methods to generate solutions. It has been a long standing conjecture that the multipole moments give a complete characterization of the stationary spacetimes. Much progress toward a proof has been made over the years. However, there is one remaining difficult task: to prove that a spacetime exists with an a-priori given arbitrary set of multipole moments subject to some given condition. Here we present such a condition for the axisymmetric case, and prove that it is both necessary and sufficient. We also extend this condition to the general case without axisymmetry, but in this case we only prove the necessity of our condition.

Humans have the ability to recognize other humans by voice alone. This is important both socially and for the robustness of speech perception. This Thesis contains a set of eight studies that investigates how different factors impact on speaker recognition and how these factors can help explain how listeners perceive and evaluate speaker identity. The first study is a review paper overviewing emotion decoding and encoding research. The second study compares the relative importance of the emotional tone in the voice and the emotional content of the message. A mismatch between these was shown to impact upon decoding speed. The third study investigates the factor dialect in speaker recognition and shows, using a bidialectal speaker as the target voice to control all other variables, that the dominance of dialect cannot be overcome. The fourth paper investigates if imitated stage dialects are as perceptually dominant as natural dialects. It was found that a professional actor could disguise his voice successfully by imitating a dialect, yet that a listener's proficiency in a language or accent can reduce susceptibility to a dialect imitation. Papers five to seven focus on automatic techniques for speaker separation. Paper five shows that a method developed for Australian English diphthongs produced comparable results with a Swedish glide + vowel transition. The sixth and seventh papers investigate a speaker separation technique developed for American English. It was found that the technique could be used to separate Swedish speakers and that it is robust against professional imitations. Paper eight investigates how age and hearing impact upon earwitness reliability. This study shows that a senior citizen with corrected hearing can be as reliable an earwitness as a younger adult with no hearing problem, but suggests that a witness' general cognitive skill deterioration needs to be considered when assessing a senior citizen's earwitness evidence. On the basis of the studies a model of speaker recognition is presented, based on the face recognition model by V. Bruce and Young (1986; British Journal of Psychology, 77, pp. 305 - 327) and the voice recognition model by Belin, Fecteau and Bédard (2004; TRENDS in Cognitive Science, 8, pp. 129 - 134). The merged and modified model handles both familiar and unfamiliar voices. The findings presented in this Thesis, in particular the findings of the individual papers in Part II, have implications for criminal cases in which speaker recognition forms a part. The findings feed directly into the growing body of forensic phonetic and forensic linguistic research.

Class numbers of algebraic number fields are central invariants. Once the underlying field has an infinite unit group they behave very irregularly due to a non-trivial regulator. This phenomenon occurs already in the simplest case of real quadratic number fields of which very little is known. Hooley derived a conjectural formula for the average of class numbers of real quadratic fields. In this thesis we extend his methods to obtain conjectural formulae and bounds for any moment, i.e., the average of an arbitrary real power of class numbers. Our formulae and bounds are based on similar (quite reasonable) assumptions of Hooley's work. In the final chapter we consider the case of the -1 power from a numerical point of view and develop an efficient algorithm to compute the average for the -1 class number power without computing class numbers.

Class numbers of algebraic number fields are central invariants. Once the underlying field has an infinite unit group they behave very irregularly due to a non-trivial regulator. This phenomenon occurs already in the simplest case of real quadratic number fields of which very little is known. Hooley derived a conjectural formula for the average of class numbers of real quadratic fields. In this thesis we extend his methods to obtain conjectural formulae and bounds for any moment, i.e., the average of an arbitrary real power of class numbers. Our formulae and bounds are based on similar (quite reasonable) assumptions of Hooley's work. In the final chapter we consider the case of the -1 power from a numerical point of view and develop an efficient algorithm to compute the average for the -1 class number power without computing class numbers.

The study of the electromagnetic phenomena along the last two centuries has brought about outstanding contributions for the human progress. The electromagnetism represents still now, at the beginning of the third millenium, a very important research area. The radiation pattern of particular types of antennas -for example, fractal or microstrip-, the analysis of the effect of the cellular communications on human beings or the detection of buried mines represent specific examples of the wide variety of problems of great interest nowadays. The study of such a variety of problems relies on the application of the Maxwell equations, which rule all the electromagnetic behaviour. Since the analytical solution can only be obtained for very particular cases of canonical forms, to tackle the analysis of an arbitrary problem, one makes use of the numerical methods. The discretization of electromagnetic integral equations by the Method of Moments -MoM- excels as a powerful and reliable tool for analysing bodies composed of locally homogeneous regions -penetrable or perfectly conducting- immerse in a wide and nearly uniform medium -typically the ground or the free-space-. These integral methods result from the surface equivalence theorem, which allows in general two different formulations, the Electric Field Integral Equation (EFIE) and the Magnetic Field Integral Equation (MFIE). For the case of penetrable bodies, the Poggio, Miller, Chang, Harrington and Wu (PMCHW) formulation, that results from the subtraction of the EFIE and MFIE at both sides of the surfaces, can also be employed.The Method of Moments is based on the full expansion of the physical magnitudes, field and current, over the interface surfaces between the regions. In consequence, the solution of the problem is obtained through the inversion of a full-matrix, which, for electrically large problems, requires excessive memory resources and computation time. That is why the MoM is widely considered a brute-force method. The expansion of the magnitudes is carried out through the discretization of the surface; that is, patches spreading over the interface. The first half of this dissertation Thesis tackles the development of the MoM applied to problems with bodies with symmetry of revolution -BoR-. Since in this case the physical magnitudes present an azimuthal periodicity, they can be expressed as a Fourier series. The orthogonality between the different modes enables to obtain separately each azimuthal mode of the solution. It is thus only required to spread the patches along the generating arc of the bodies for each mode, which is very advantageous because the electromagnetic analysis can be carried out indeed for dimensionally large problems. A well-known PeC-EFIE BoR formulation is developed. Accordingly, PeC-MFIE and PMCHW formulations are developed from scratch. Furthermore, it is commented in detail and corrected to some extent the numerical error associated to the fastest-varying part of the PeC-MFIE BoR operator. The BoR-codes are particularly useful in modelling the electromagnetic behaviour of buried mines, which very often show revolution symmetry. The most outstanding contribution of this dissertation Thesis is the study of the appropriate conditions to develop correctly the 3D operators so as to yield accurate results for any structure. Since the discretization implies a break on the continuity properties of the physical magnitudes -field and current- the valid 3D-operators must ensure the physical electromagnetic requirements in the discretized surface. In mathematical terms, these requirements set the rank -field- and domain- -current- spaces, which essentially require the enforcement of the continuity across the edges of either the tangential or the normal component of the expanded magnitudes.For the case of an arbitrary perfectly conducting -PeC- body, it is recommended in this work the use of the divergence-conforming and of the curl-conforming functions respectively in the development of the PeC-EFIE and the PeC-MFIE operators. Low-order sets over triangular facets -RWG and unxRWG- are chosen to develop the PeC-operators. Furthermore, it is reasoned theoretically the inherent misbehaviour of the PeC-MFIE in case the current expansion relies on a divergence-conforming set. A heuristic correction is provided. The better behaviour of PeC-EFIE(RWG) and PeC-MFIE(unxRWG) is confirmed with examples. In view of the results, it is reasoned the suitability of PeC-EFIE(RWG) for the analysis of physical polyhedrons, which makes PeC-MFIE(unxRWG) excel as a more appropriate operator for curved bodies. A procedure for improving the performance of PeC-EFIE(RWG) for coarsely meshed spheres is given.For the case of arbitrary penetrable bodies, the same low-order sets are used to expand the operators EFIE, MFIE and PMCHW. It is shown their compatibility with the combination of the right PeC-operators. In the dielectric case, in addition to the required continuity of the magnitudes across the edges at each region, the fields at both sides of the surface must satisfy the interface continuity, which is ignored in the conducting case -the fields are null inside the conductor-. The impossibility of meeting both continuity requirements at the same time justifies the apparition of inherent and different errors in the dual EFIE-MFIE and in PMCHW. It is thoroughly reasoned and confirmed with examples the suitability of PMCHW for problems with only penetrable regions. It is also shown and discussed in detail the robustness of EFIE-MFIE since its behaviour is appropriate for electrically not too small structures with perfectly conducting or penetrable regions. The analysis of composite structures -very useful to model microstrip antennas- can be considered as a group of disjoint bodies with null distances of separation. For this type of problems, it is recommended in this work the use of EFIE-MFIE since, unlike PMCHW, they can ensure the continuous transition to zero of a distance of separation increasingly small. Finally, efficient methods -IE-MEI and MLFMM- relying on the previous 3D-operators. The development of the PeC 3D IE-MEI cannot maintain the advantages present in the 2D case since the harmonic metrons are not valid in the 3D general case. A new set of metrons that ensures little discontinuity of the current across the edges is presented. It is confirmed with examples how these metrons, so-called quasi-continuous, reduce the number of required coefficients per row for a certain current error. Some examples of penetrable spheres with moderate electrical dimensions analysed under a MLFMM implementation are shown and commented.

Dans cette thèse, je propose une généralisation de la méthode de vraisemblance empirique généralisée (GEL) pour permettre la possibilité d'avoir soit un très grand nombre de conditions de moment ou des conditions définies sur un continuum. Cette généralisation peut permettre par exemple d'estimer des modèles de régression avec régresseurs endogènes pour lesquels le nombre d'instruments est très élevé ou encore que la relation entre les régresseurs et les variables exogènes est inconnue. Il est également possible de baser notre estimation sur des conditions de moment construites à partir de fonctions caractéristiques. Il devient alors possible d'estimer les coefficients d'une distribution quelconque ou d'un processus stochastique lorsque sa fonction de vraisemblance n'admet pas de forme analytique. C'est le cas entre autres de la distribution stable et de la plupart des processus de diffusion exprimés en temps continu. Cette généralisation a été proposée par (Carrasco and Florens, 2000) pour la méthode des moments généralisés (CGMM). Sur la base des résultats de (Newey and Smith, 2004), qui démontrent la supériorité asymptotique de GEL sur GMM, la méthode que je propose représente donc une contribution substantielle. La thèse est divisée en trois chapitres. Le premier présente en détails la méthode de vraisemblance empirique généralisée pour un continuum de moments (CGEL), démontre la convergence en probabilité et en distribution de ses estimateurs et décrit la procédure à suivre en pratique pour estimer les coefficients du modèle à l'aide d'une approche matricielle relativement simple. De plus, je démontre l'équivalence asymptotique de CGEL et CGMM. CGEL est en fait un algorithme non-linéaire régularisé à la Tikhonov, qui permet d'obtenir l'estimateur GEL dans le cas où le nombre de conditions est très grand. Dans cette méthode, un paramètre de régularisation, αn, permet de résoudre le problème d'optimisation mal posé qui en résulte et d'obtenir une solution unique et stable. Le paramètre αn doit converger vers zéro lentement lorsque la taille d'échantillon augmente pour que l'estimateur soit convergent et que la solution demeure stable. Les détails du rythme de convergence de αn sont également présentés dans ce chapitre. Finalement, le chapitre présente la façon de tester les conditions de moments en généralisant les trois tests de spécifications existants pour GEL. Dans le chapitre 2, je présente plusieurs applications numériques. L'objectif est de voir les possibilités de CGEL, d'analyser les propriétés et ses estimateurs en échantillons finis, en comparaison avec ceux de CGMM, et de comprendre l'impact du paramètre αn sur le biais et la variance des estimateurs. Les applications analysées sont : l'estimation d'un modèle linéaire avec endogénéité de forme inconnue, l'estimation des paramètres d'une distribution stable et l'estimation des coefficients d'un processus de diffusion. De façon générale les résultats démontrent que la dominance de CGEL sur CGMM dépend de la valeur de αn. Cela démontre en fait la nécessité de développer une méthode de sélection de αn. Finalement, une méthode de sélection du paramètre an est proposée dans le dernier chapitre. Dans un premier temps, je démontre qu'une méthode de bootstrap simple permet difficilement de faire un choix optimal car elle produit une relation très volatile entre αn et l'erreur quadratique moyen (MSE) du coefficient. Ensuite, je présente une approximation de second ordre du MSE de CGEL par un développement stochastique des conditions de premier ordre comme fait par (Donald and Newey, 2001) pour les double moindres carrés, (Donald, Imbens and Newey, 2010) pour GEL ainsi que (Carrasco, 2010) et (Carrasco and Kotchoni, 2010) pour CGMM. Cette approche permet d'obtenir une relation lisse entre αn et le MSE et donc d'utiliser un algorithme d'optimisation pour obtenir le paramètre optimal. Les résultats semblent être conformes aux résultants précédents selon lesquels la méthode de vraisemblance empirique domine les autres méthodes faisant partie de la famille CGEL. Ils semblent également suggérer que αn, pour le cas linéaire considéré, devrait être choisi aussi petit que possible car c'est de cette façon que le MSE est minimisé. ______________________________________________________________________________ MOTS-CLÉS DE L’AUTEUR : Vraisemblance Généralisée, Continuum de moments, Méthode des moments généralisés, Économétrie, Variables Instrumentales