The classical moment problem concerns distribution functions on the real line. The central feature is the connection between distribution functions and the moment sequences which they generate via a Stieltjes integral. The solution of the classical moment problem leads to the well known theorem of Favard which connects orthogonal polynomial sequences with distribution functions on the real line. Orthogonal polynomials in their turn arise in the computation of measures via continued fractions and the Nevanlinna parametrisation. In this dissertation classical orthogonal polynomials are investigated rst and their connection with hypergeometric series is exhibited. Results from the moment problem allow the study of a more general class of orthogonal polynomials. q-Hypergeometric series are presented in analogy with the ordinary hypergeometric series and some results on q-Laguerre polynomials are given. Finally recent research will be discussed.

The use of a stochastic model to predict the likelihood of future outcomes forms an integral part of decision optimization under uncertainty. In classical stochastic modeling uncertain parameters are often assumed to be driven by a particular form of probability distribution. In practice however, the distributional form is often difficult to infer from the observed data, and the incorrect choice of distribution can lead to significant quality deterioration of resultant decisions and unexpected losses. In this thesis, we present new approaches for evaluating expected future performance that do not rely on an exact distributional specification and can be robust against the errors related to committing to a particular specification. The notion of comprehensive robustness is promoted, where various degrees of model misspecification are studied. This includes fundamental one such as unknown distributional form and more involved ones such as stochastic moments and moment outliers. The approaches are developed based on the techniques of moment-based optimization, where bounds on the expected performance are sought based solely on partial moment information. They can be integrated into decision optimization and generate decisions that are robust against model misspecification in a comprehensive manner. In the first part of the thesis, we extend the applicability of moment-based optimization to incorporate new objective functions such as convex risk measures and richer moment information such as higher-order multivariate moments. In the second part, new tractable optimization frameworks are developed that account for various forms of moment uncertainty in the context of decision analysis and optimization. Financial applications such as portfolio selection and option pricing are studied.

The use of a stochastic model to predict the likelihood of future outcomes forms an integral part of decision optimization under uncertainty. In classical stochastic modeling uncertain parameters are often assumed to be driven by a particular form of probability distribution. In practice however, the distributional form is often difficult to infer from the observed data, and the incorrect choice of distribution can lead to significant quality deterioration of resultant decisions and unexpected losses. In this thesis, we present new approaches for evaluating expected future performance that do not rely on an exact distributional specification and can be robust against the errors related to committing to a particular specification. The notion of comprehensive robustness is promoted, where various degrees of model misspecification are studied. This includes fundamental one such as unknown distributional form and more involved ones such as stochastic moments and moment outliers. The approaches are developed based on the techniques of moment-based optimization, where bounds on the expected performance are sought based solely on partial moment information. They can be integrated into decision optimization and generate decisions that are robust against model misspecification in a comprehensive manner. In the first part of the thesis, we extend the applicability of moment-based optimization to incorporate new objective functions such as convex risk measures and richer moment information such as higher-order multivariate moments. In the second part, new tractable optimization frameworks are developed that account for various forms of moment uncertainty in the context of decision analysis and optimization. Financial applications such as portfolio selection and option pricing are studied.

Thesis (Ph. D.)--University of Missouri-Columbia, 2001. / Typescript. Vita. Includes bibliographical references (leaves 51-53). Also available on the Internet.

Thesis (Ph. D.)--University of Missouri-Columbia, 2001. / Typescript. Vita. Includes bibliographical references (leaves 51-53). Also available on the Internet.

Using recent progress on moment problems, and their connections with semidefinite optimization, we present in this paper a new methodology based on semidefinite optimization, to obtain a hierarchy of upper and lower bounds on both linear and certain nonlinear functionals defined on solutions of linear partial differential equations. We apply the proposed methods to examples of PDEs in one and two dimensions with very encouraging results. We also provide computation evidence that the semidefinite constraints are critically important in improving the quality of the bounds, that is without them the bounds are weak. / Singapore-MIT Alliance (SMA)

Die vorliegende Arbeit beschäftigt sich mit den finiten matriziellen Momentenproblemen von Stieltjes-Typ und beschreibt unter Verwendung der Methode der Fundamentalen Matrixungleichungen die Lösungsmenge durch gebrochen lineare Transformationen.

In this dissertation, we first present a unified treatment of compact moment problems, both the truncated and full moment cases. Second, we define the lower and upper functions V±(ð₁,... ð _n) on the convex hull of the curve Γ_n = {(t,.·.,tⁿ): t ∈ [0,1] } for each positive integer n. Explicit formulas of these functions are derived and applied to the study of the subnormal completion problem in operator theory. Last, we show that certain power functions are the building blocks of completely positive functions; by our definition, these functions are the continuous functions on the interval [0, 1] that map each Hausdorff moment sequence of a probability measure into another one. / Ph. D.

The present thesis intents on analysing the truncated matricial Hamburger power moment problem in the general (degenerate and non-degenerate) case. Initiated due to manifold lines of research, by this time, outnumbering results and thoughts have been established that are concerned with specific subproblems within this field. The resulting presence of such a diversity as well as an extensively considered topic si- multaneously involves advantageous as well as obstructive aspects: on the one hand, we adopt the favourable possibility to capitalise on essential available results that proved beneficial within subsequent research. Nevertheless, on the other hand, we are obliged to illustrate major preparatory work in order to illucidate the comprehension of the attaching examination. Moreover, treating the matricial cases of the respective problems requires meticulous technical demands, in particular, in view of the chosen explicit approach to solving the considered tasks. Consequently, the first part of this thesis is dedicated to furnishing the necessary basis arranging the prime results of this research paper. Compul- sary notation as well as objects are introduced and thoroughly explained. Furthermore, the required techniques in order to achieve the desired results are characterised and ex- haustively discussed. Concerning the respective findings, we are afforded the opportunity to seise presentations and results that are, by this time, elaborately studied. Being equipped with mandatory cognisance, the thematically bipartite second and pivo- tal part objectives to describe all the possible values of all the solution functions of the truncated matricial Hamburger power moment problem M P [R; (s j ) 2n j=0 , ≤]. Aming this, we realise a first paramount achievement epitomising one of the two parts of the main results: Capturing an established representation of the solution set R 0,q [Π + ; (s j ) 2n j=0 , ≤] of the assigned matricial Hamburger moment problem via operating a specific algorithm of Schur-type, we expand these findings. We formulate a parameterisation of the set R 0,q [Π + ; (s j ) 2n j=0 , ≤] which is compatible with establishing respective equivalence classes within a certain subset of Nevanlinna pairs and utilise specific systems of orthogonal polynomials in order to entrench novel representations. In conclusion, we receive a para- meterisation that is valid within the entire upper open complex half-plane Π + . The second of the two prime parts changes focus to analysing all possible values of the functions belonging to R 0,q [Π + ; (s j ) 2n j=0 , ≤] in an arbitrary point w ∈ Π + . We gain two decisive conclusions: We identify these respective values to exhaust particular matrix balls 2n K[(s j ) 2n j=0 , w] := {F (w) | F ∈ R 0,q [Π + ; (s j ) j=0 , ≤]} the parameters of which are feasable to being described by specific rational matrix-valued functions and, in this course, enhance formerly established analyses. Moreover, we compile an alternative representation of the semi-radii constructing the respective matrix balls which manifests supportive in further consideration. We seise the achieved parameterisation of the set K[(s j ) 2n j=0 , w] and examine the behaviour of the respective sequences of left and right semi-radii. We recognise that these sequences of semi-radii associated with the respective matrix balls in the general case admit a particular monotonic behaviour. Consequently, with increasing number of given data, the resulting matrix balls are identified as being nested. Moreover, a proper description of the limit case of an infinite number of prescribed moments is facilitated.:1. Brief Historic Embedding and Introduction 2. Part I: Initialising Compulsary Cognisance Arranging Principal Achievements 2.1. Notation and Preliminaries 2.2. Particular Classes of Holomorphic Matrix-Valued Functions 2.3. Nevanlinna Pairs 2.4. Block Hankel Matrices 2.5. A Schur-Type Algorithm for Sequences of Complex p × q Matrices 2.6. Specific Matrix Polynomials 3. Part II: Momentous Results and Exposition – Improved Parameterisations of the Set R 0,q [Π + ; (s j ) 2n j=0 , ≤] 3.1. An Essential Step to a Parameterisation of the Solution Set R 0,q [Π + ; (s j ) 2n j=0 , ≤] 3.2. Parameterisation of the Solution Set R 0,q [Π + ; (s j ) 2n j=0 3.3. Particular Matrix Polynomials 3.4. Description of the Solution Set of the Truncated Matricial Hamburger Moment Problem by a Certain System of Orthogonal Matrix Polynomials 4. Part III: Prime Results and Exposition – Novel Description Balls 4.1. Particular Rational Matrix-Valued Functions 4.2. Description of the Values of the Solutions 4.3. Monotony of the Semi-Radii and Limit Balls of the Weyl Matrix 5. Summary of Principal Achievements and Prospects A. Matrix Theory B. Integration Theory of Non-Negative Hermitian Measures

Die vorliegende Arbeit beschäftigt sich mit den finiten matriziellen Momentenproblemen von Stieltjes-Typ und beschreibt unter Verwendung der Methode der Fundamentalen Matrixungleichungen die Lösungsmenge durch gebrochen lineare Transformationen.