published_or_final_version / Civil and Structural Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / Statistics and Actuarial Science / Doctoral / Doctor of Philosophy

published_or_final_version / abstract / Industrial and Manufacturing Systems Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / abstract / Mathematics / Master / Master of Philosophy

published_or_final_version / Mechanical Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / Statistics and Actuarial Science / Master / Master of Philosophy

摘要 本文於確定給付退休金計劃下，探討基金經理人於最差基金財務短絀情境發生前極大化管理目標之最適投資組合，基金比值過程定義為基金現值與負債指標之比例，管理人將於指定最差基金比值發生前極大化達成既定經營目標之機率，隨時間改變之基金投資集合包括無風險之現金、債券與股票。本研究建構隨機控制模型描述此最適化問題，並以動態規劃方法求解，由結果歸納，經理人之最適策略包含極小化基金比值變異之避險因素，風險偏好及跨期投資集合相關之避險因素與模型狀態變數相關之避險因素。本研究利用馬可夫練逼近法逼近隨機控制的數值解，結果顯示基金經理人須握有很大部位的債券，且不同的投資期間對於最適投資決策有很大的影響。 關鍵字: 短絀、確定給付、負債指標、隨機控制、動態規劃。 / Abstract This paper analyzes the portfolio problem that is a pension fund manager has to maximize the possibility of reaching his managerial goal before the worst scenario shortfall occurs in a defined benefit pension scheme. The fund ratio process defined as the ratio between the fund level and its accrued liability benchmark is attained to maximize the probability that the predetermined target is achieved before it falls below an intolerable boundary. The time-varying opportunity set in our study includes risk-free cash, bonds and stock index. The problems are formulated as a stochastic control framework and are solved through dynamics programming. In this study, the optimal portfolio are characterized by three components, the liability hedging component, the intertemporal hedging component against changes in the opportunity set, and the temporal hedging component minimizing the variation in fund ratio growth. The Markov chain approximation methods are employed to approximate the stochastic control solutions numerically. The result shows that fund managers should hold large proportions of bonds and time horizon plays a crucial role in constructing the optimal portfolio. Keywords: shortfall; defined benefit; liability benchmark; stochastic control; dynamic programming.

A stochastic dynamic programming model (called P.B.S.D.P.) based on the consideration of peak discharge and time between peaks as two stochastic variables has been used to model and to solve a reservoir operation problem. This conceptualization of the physical reality allows to solve, in this order, the tactical and strategic operation of surface reservoirs. This P.B.S.D.P. model has been applied to the Sau reservoir in the Northeastern corner of Spain. The results showed a significant improvement over the currently used operation procedure, yielding values of yearly average electricity production that are somewhat under 6% of what could have been the maximum electricity production.

This dissertation develops an analytical-numerical approach to deterministically predict the space-time evolution of concentrations in heterogeneous geologic media conditioned on measurements of hydraulic conductivities (transmissivities) and/or hydraulic heads. Based on the new conditional Eulerian-Lagrangian transport theory by Neuman, we solve the conditional transport problem analytically at early time, and express it in pseudo-Fickian form at late time. The stochastically derived deterministic pseudo-Fickian mean concentration equation involves a conditional, space-time dependent dispersion tensor. The latter not only depends on properties of the medium and the velocity but also on the available information, and can be evaluated numerically along mean "particle" trajectories. The transport equation lends itself to accurate solution by standard Galerkin finite elements on a relatively coarse grid. This approach allows computing without using Monte Carlo simulation and explicitly the following: Concentration variance/covariance (uncertainty), origin of detected contaminant and associated uncertainty, mass flow rate across a "compliance surface", cumulative mass release and travel time probability distribution across this surface, uncertainty associated with the latter, second spatial moment of conditional mean plume about its center of mass, conditional mean second spatial moment of actual plume about its center of mass, conditional co-variance of plume center of mass, and effect of non-Gaussian velocity distribution. This approach can also account for uncertainty in initial mass and/or concentration when predicting the future evolution of a plume, whereas almost all existing stochastic models of solute transport assume the initial state to be known with certainty. We illustrate this approach by considering deterministic and uncertain instantaneous point and nonpoint sources in a two-dimensional domain with a mildly fluctuating, statistically homogeneous, lognormal transmissivity field. We take the unconditional mean velocity to be uniform, but allow conditioning on log transmissivity and hydraulic head data. Conditioning renders the velocity field statistically nonhomogeneous with reduced variances and correlation scales, renders the predicted plume irregular and non-Gaussian, and generally reduces both predictive dispersion and uncertainty.

Maximum-likelihood estimation or other probabilistic estimation methods are underused in many areas of applied gamma-ray imaging, particularly in biomedicine. In this work, we show how to use our understanding of stochastic processes in a scintillation camera and their effect on signal formation to better estimate gamma-ray interaction parameters such as interaction position or energy.To apply statistical estimation methods, we need an accurate description of the signal statistics as a function of the parameters to be estimated. First, we develop a probability model of the signals conditioned on the parameters to be estimated by carefully examining the signal generation process. Subsequently, the likelihood model is calibrated by measuring signal statistics for an ensemble of events as a function of the estimate parameters.In this work, we investigate the application of ML-estimation methods for three topics. First, we design, build, and evaluate a scintillation camera based on a multi-anode PMT readout for use with ML-estimation techniques. Next, we develop methods for calibrating the response statistics of a thick-detector gamma camera as a function of interaction depth. Finally, we demonstrate the use of ML estimation with a modified clinical Anger camera.