Adaptive designs in the time to event setting: The potential for benefit and risk

Thesis (Ph.D.)--University of Washington, 2016-06

Evaluating trends in abundance of immature green turtles, Chelonia mydas, in the Greater Caribbean

Many long-lived marine species exhibit life history traits. that make them more vulnerable to overexploitation. Accurate population trend analysis is essential for development and assessment of management plans for these species. However, because many of these species disperse over large geographic areas, have life stages inaccessible to human surveyors, and/or undergo complex developmental migrations, data on trends in abundance are often available for only one stage of the population, usually breeding adults. The green turtle (Chelonia mydas) is one of these long-lived species for which population trends are based almost exclusively on either numbers of females that emerge to nest or numbers of nests deposited each year on geographically restricted beaches. In this study, we generated estimates of annual abundance for juvenile green turtles at two foraging grounds in the Bahamas based on long-term capture-mark-recapture (CMR) studies at Union Creek (24 years) and Conception Creek (13 years), using a two-stage approach. First, we estimated recapture probabilities from CMR data using the Cormack-Jolly-Seber models in the software program MARK; second, we estimated annual abundance of green turtles. at both study sites using the recapture probabilities in a Horvitz-Thompson type estimation procedure. Green turtle abundance did not change significantly in Conception Creek, but, in Union Creek, green turtle abundance had successive phases of significant increase, significant decrease, and stability. These changes in abundance resulted from changes in immigration, not survival or emigration. The trends in abundance on the foraging grounds did not conform to the significantly increasing trend for the major nesting population at Tortuguero, Costa Rica. This disparity highlights the challenges of assessing population-wide trends of green turtles and other long-lived species. The best approach for monitoring population trends may be a combination of (1) extensive surveys to provide data for large-scale trends in relative population abundance, and (2) intensive surveys, using CMR techniques, to estimate absolute abundance and evaluate the demographic processes' driving the trends.

Interocular masking and summation indicate two stages of divisive contrast gain control

Our understanding of early spatial vision owes much to contrast masking and summation paradigms. In particular, the deep region of facilitation at low mask contrasts is thought to indicate a rapidly accelerating contrast transducer (eg a square-law or greater). In experiment 1, we tapped an early stage of this process by measuring monocular and binocular thresholds for patches of 1 cycle deg-1 sine-wave grating. Threshold ratios were around 1.7, implying a nearly linear transducer with an exponent around 1.3. With this form of transducer, two previous models (Legge, 1984 Vision Research 24 385 - 394; Meese et al, 2004 Perception 33 Supplement, 41) failed to fit the monocular, binocular, and dichoptic masking functions measured in experiment 2. However, a new model with two-stages of divisive gain control fits the data very well. Stage 1 incorporates nearly linear monocular transducers (to account for the high level of binocular summation and slight dichoptic facilitation), and monocular and interocular suppression (to fit the profound 42 Oral presentations: Spatial vision Thursday dichoptic masking). Stage 2 incorporates steeply accelerating transduction (to fit the deep regions of monocular and binocular facilitation), and binocular summation and suppression (to fit the monocular and binocular masking). With all model parameters fixed from the discrimination thresholds, we examined the slopes of the psychometric functions. The monocular and binocular slopes were steep (Weibull ߘ3-4) at very low mask contrasts and shallow (ߘ1.2) at all higher contrasts, as predicted by all three models. The dichoptic slopes were steep (ߘ3-4) at very low contrasts, and very steep (ß>5.5) at high contrasts (confirming Meese et al, loco cit.). A crucial new result was that intermediate dichoptic mask contrasts produced shallow slopes (ߘ2). Only the two-stage model predicted the observed pattern of slope variation, so providing good empirical support for a two-stage process of binocular contrast transduction. [Supported by EPSRC GR/S74515/01]

Purification of plasmid DNA for use in human gene therapy

Two key issues defined the focus of this research in manufacturing plasmid DNA for use In human gene therapy. First, the processing of E.coli bacterial cells to effect the separation of therapeutic plasmid DNA from cellular debris and adventitious material. Second, the affinity purification of the plasmid DNA in a Simple one-stage process. The need arises when considering the concerns that have been recently voiced by the FDA concerning the scalability and reproducibility of the current manufacturing processes in meeting the quality criteria of purity, potency, efficacy, and safety for a recombinant drug substance for use in humans. To develop a preliminary purification procedure, an EFD cross-flow micro-filtration module was assessed for its ability to effect the 20-fold concentration, 6-time diafiltration, and final clarification of the plasmid DNA from the subsequent cell lysate that is derived from a 1 liter E.coli bacterial cell culture. Historically, the employment of cross-flow filtration modules within procedures for harvesting cells from bacterial cultures have failed to reach the required standards dictated by existing continuous centrifuge technologies, frequently resulting in the rapid blinding of the membrane with bacterial cells that substantially reduces the permeate flux. By challenging the EFD module, containing six helical wound tubular membranes promoting centrifugal instabilities known as Dean vortices, with distilled water between the Dean number's of 187Dn and 818Dn,and the transmembrane pressures (TMP) of 0 to 5 psi. The data demonstrated that the fluid dynamics significantly influenced the permeation rate, displaying a maximum at 227Dn (312 Imh) and minimum at 818Dn (130 Imh) for a transmembrane pressure of 1 psi. Numerical studies indicated that the initial increase and subsequent decrease resulted from a competition between the centrifugal and viscous forces that create the Dean vortices. At Dean numbers between 187Dn and 227Dn , the forces combine constructively to increase the apparent strength and influence of the Dean vortices. However, as the Dean number in increases above 227 On the centrifugal force dominates the viscous forces, compressing the Dean vortices into the membrane walls and reducing their influence on the radial transmembrane pressure i.e. the permeate flux reduced. When investigating the action of the Dean vortices in controlling tile fouling rate of E.coli bacterial cells, it was demonstrated that the optimum cross-flow rate at which to effect the concentration of a bacterial cell culture was 579Dn and 3 psi TMP, processing in excess of 400 Imh for 20 minutes (i.e., concentrating a 1L culture to 50 ml in 10 minutes at an average of 450 Imh). The data demonstrated that there was a conflict between the Dean number at which the shear rate could control the cell fouling, and the Dean number at which tile optimum flux enhancement was found. Hence, the internal geometry of the EFD module was shown to sub-optimal for this application. At 579Dn and 3 psi TMP, the 6-fold diafiltration was shown to occupy 3.6 minutes of process time, processing at an average flux of 400 Imh. Again, at 579Dn and 3 psi TMP the clarification of the plasmid from tile resulting freeze-thaw cell lysate was achieved at 120 Iml1, passing 83% (2,5 mg) of the plasmid DNA (6,3 ng μ-1 10.8 mg of genomic DNA (∼23,00 Obp, 36 ng μ-1 ), and 7.2 mg of cellular proteins (5-100 kDa, 21.4 ngμ-1 ) into the post-EFD process stream. Hence the EFD module was shown to be effective, achieving the desired objectives in approximately 25 minutes. On the basis of its ability to intercalate into low molecular weight dsDNA present in dilute cell lysates, and be electrophoresed through agarose, the fluorophore PicoGreen was selected for the development of a suitable dsDNA assay. It was assesseel for its accuracy, and reliability, In determining the concentration and identity of DNA present in samples that were eleclrophoresed through agarose gels. The signal emitted by intercalated PicoGreen was shown to be constant and linear, and that the mobility of the PicaGreen-DNA complex was not affected by the intercalation. Concerning the secondary purification procedure, various anion-exchange membranes were assessed for their ability to capture plasmid DNA from the post-EFD process stream. For a commercially available Sartorius Sartobind Q15 membrane, the reduction in the equilibriumbinding capacity for  ctDNA in buffer of increasing ionic demonstrated that DNA was being.adsorbed by electrostatic  interactions only. However, the problems associated with fluid distribution across the membrane demonstrated that the membrane housing was the predominant cause of the .erratic breakthrough curves. Consequently, this would need to be rectified before such a membrane could be integrated into the current system, or indeed be scaled beyond laboratory scale. However, when challenged with the process material, the data showed that considerable quantities of protein (1150 μg) were adsorbed preferentially to the plasmid DNA (44 μg). This was also shown for derived Pall Gelman UltraBind US450 membranes that had been functionalised by varying molecular weight poly-L~lysine and polyethyleneimine ligands. Hence the anion-exchange membranes were shown to be ineffective in capturing plasmid DNA from the process stream. Finally, work was performed to integrate a sequence-specific DNA·binding protein into a single-stage DNA chromatography, isolating plasmid DNA from E.coli cells whilst minimising the contamination from genomic DNA and cellular protein. Preliminary work demonstrated that the fusion protein was capable of isolating pUC19 DNA into which the recognition sequence for the fusion-protein had been inserted (pTS DNA) when in the presence of the conditioned process material. Althougth the pTS recognition sequence differs from native pUC19 sequences by only 2 bp, the fusion protein was shown to act as a highly selective affinity ligand for pTS DNA alone. Subsequently, the scale of the process was scaled 25-fold and positioned directly following the EFD system. In conclusion, the integration of the EFD micro-filtration system and zinc-finger affinity purification technique resulted in the capture of approximately 1 mg of plasmid DNA was purified from 1L of E.coli  culture in a simple two stage process, resulting in the complete removal of genomic DNA and 96.7% of cellular protein in less than 1 hour of process time.

The development of chemical exposure limits for the workplace

The thesis examines and explains the development of occupational exposure limits (OELs) as a means of preventing work related disease and ill health. The research focuses on the USA and UK and sets the work within a certain historical and social context. A subsidiary aim of the thesis is to identify any short comings in OELs and the methods by which they are set and suggest alternatives. The research framework uses Thomas Kuhn's idea of science progressing by means of paradigms which he describes at one point, `lq ... universally recognised scientific achievements that for a time provide model problems and solutions to a community of practitioners. KUHN (1970). Once learned individuals in the community, `lq ... are committed to the same rules and standards for scientific practice. Ibid. Kuhn's ideas are adapted by combining them with a view of industrial hygiene as an applied science-based profession having many of the qualities of non-scientific professions. The great advantage of this approach to OELs is that it keeps the analysis grounded in the behaviour and priorities of the groups which have forged, propounded, used, benefited from, and defended, them. The development and use of OELs on a larger scale is shown to be connected to the growth of a new profession in the USA; industrial hygiene, with the assistance of another new profession; industrial toxicology. The origins of these professions, particularly industrial hygiene, are traced. By examining the growth of the professions and the writings of key individuals it is possible to show how technical, economic and social factors became embedded in the OEL paradigm which industrial hygienists and toxicologists forged. The origin, mission and needs of these professions and their clients made such influences almost inevitable. The use of the OEL paradigm in practice is examined by an analysis of the process of the American Conference of Governmental Industrial Hygienists, Threshold Limit Value (ACGIH, TLV) Committee via the Minutes from 1962-1984. A similar approach is taken with the development of OELs in the UK. Although the form and definition of TLVs has encouraged the belief that they are health-based OELs the conclusion is that they, and most other OELs, are, and always have been, reasonably practicable limits: the degree of risk posed by a substance is weighed against the feasibility and cost of controlling exposure to that substance. The confusion over the status of TLVs and other OELs is seen to be a confusion at the heart of the OEL paradigm and the historical perspective explains why this should be. The paradigm has prevented the creation of truly health-based and, conversely, truly reasonably practicable OELs. In the final part of the thesis the analysis of the development of OELs is set in a contemporary context and a proposal for a two-stage, two-committee procedure for producing sets of OELs is put forward. This approach is set within an alternative OEL paradigm. The advantages, benefits and likely obstacles to these proposals are discussed.

Optical regeneration at 40 Gb/s in dispersion-managed transmission lines with in-line synchronous modulators

Numerical optimization is performed of the 40-Gb/s dispersion-managed (DM) soliton transmission system with in-line synchronous intensity modulation. Stability of DM soliton transmission results from a combined action of dispersion, nonlinearity, in-line filtering, and modulation through effective periodic bandwidth management of carrier pulses. Therefore, analysis of the multiparametric problem is typically required. A two-stage time-saving numerical optimization procedure is applied. At the first step, the regions of the stable carrier propagation are determined using theoretical models available for DM solitons, and system parameters are optimized. At the second stage, full numerical simulations are undertaken in order to verify the tolerance of optimal transmission regimes. An approach developed demonstrates feasibility of error-free transmission over 20 000 km in a transmission line composed of standard fiber and dispersion compensation fiber at 40 Gb/s.

Land resource requirements for bioenergy in India

This study presents a two stage process to determine suitable areas to grow fuel crops: i) FAO Agro Ecological Zones (AEZ) procedure is applied to four Indian states of different geographical characteristics; and ii) Modelling the growth of candidate crops with GEPIC water and nutrient model, which is used to determine potential yield of candidate crops in areas where irrigation water is brackish or soil is saline. Absence of digital soil maps, paucity of readily available climate data and knowledge of detailed requirements of candidate crops are some of the major problems, of which, a series of detailed maps will evaluate true potential of biofuels in India.

Cooperation under incomplete contracting

We examine the notion of the core when cooperation takes place in a setting with time and uncertainty. We do so in a two-period general equilibrium setting with incomplete markets. Market incompleteness implies that players cannot make all possible binding commitments regarding their actions at different date-events. We unify various treatments of dynamic core concepts existing in the literature. This results in definitions of the Classical Core, the Segregated Core, the Two-stage Core, the Strong Sequential Core, and the Weak Sequential Core. Except for the Classical Core, all these concepts can be defined by requiring absence of blocking in period 0 and at any date-event in period 1. The concepts only differ with respect to the notion of blocking in period 0. To evaluate these concepts, we study three market structures in detail: strongly complete markets, incomplete markets in finance economies, and incomplete markets in settings with multiple commodities.

Stopping rules in Bayesian adaptive threshold estimation

Threshold estimation with sequential procedures is justifiable on the surmise that the index used in the so-called dynamic stopping rule has diagnostic value for identifying when an accurate estimate has been obtained. The performance of five types of Bayesian sequential procedure was compared here to that of an analogous fixed-length procedure. Indices for use in sequential procedures were: (1) the width of the Bayesian probability interval, (2) the posterior standard deviation, (3) the absolute change, (4) the average change, and (5) the number of sign fluctuations. A simulation study was carried out to evaluate which index renders estimates with less bias and smaller standard error at lower cost (i.e. lower average number of trials to completion), in both yes–no and two-alternative forced-choice (2AFC) tasks. We also considered the effect of the form and parameters of the psychometric function and its similarity with themodel function assumed in the procedure. Our results show that sequential procedures do not outperform fixed-length procedures in yes–no tasks. However, in 2AFC tasks, sequential procedures not based on sign fluctuations all yield minimally better estimates than fixed-length procedures, although most of the improvement occurs with short runs that render undependable estimates and the differences vanish when the procedures run for a number of trials (around 70) that ensures dependability. Thus, none of the indices considered here (some of which are widespread) has the diagnostic value that would justify its use. In addition, difficulties of implementation make sequential procedures unfit as alternatives to fixed-length procedures.

SINGLE STAGE LED DRIVER TECHNOLOGY WITH LOW FREQUENCY RIPPLE CANCELLATION

Because of high efficacy, long lifespan, and environment-friendly operation, LED lighting devices become more and more popular in every part of our life, such as ornament/interior lighting, outdoor lightings and flood lighting. The LED driver is the most critical part of the LED lighting fixture. It heavily affects the purchasing cost, operation cost as well as the light quality. Design a high efficiency, low component cost and flicker-free LED driver is the goal. The conventional single-stage LED driver can achieve low cost and high efficiency. However, it inevitably produces significant twice-line-frequency lighting flicker, which adversely affects our health. The conventional two-stage LED driver can achieve flicker-free LED driving at the expenses of significantly adding component cost, design complexity and low the efficiency. The basic ripple cancellation LED driving method has been proposed in chapter three. It achieves a high efficiency and a low component cost as the single-stage LED driver while also obtaining flicker-free LED driving performance. The basic ripple cancellation LED driver is the foundation of the entire thesis. As the research evolving, another two ripple cancellation LED drivers has been developed to improve different aspects of the basic ripple cancellation LED driver design. The primary side controlled ripple cancellation LED driver has been proposed in chapter four to further reduce cost on the control circuit. It eliminates secondary side compensation circuit and an opto-coupler in design while at the same time maintaining flicker-free LED driving. A potential integrated primary side controller can be designed based on the proposed LED driving method. The energy channeling ripple cancellation LED driver has been proposed in chapter five to further reduce cost on the power stage circuit. In previous two ripple cancellation LED drivers, an additional DC-DC converter is needed to achieve ripple cancellation. A power transistor has been used in the energy channeling ripple cancellation LED driving design to successfully replace a separate DC-DC converter and therefore achieved lower cost. The detailed analysis supports the theory of the proposed ripple cancellation LED drivers. Simulation and experiment have also been included to verify the proposed ripple cancellation LED drivers.

Optimal static and sequential design : a critical review

The aim of this thesis is to review and augment the theory and methods of optimal experimental design. In Chapter I the scene is set by considering the possible aims of an experimenter prior to an experiment, the statistical methods one might use to achieve those aims and how experimental design might aid this procedure. It is indicated that, given a criterion for design, a priori optimal design will only be possible in certain instances and, otherwise, some form of sequential procedure would seem to be indicated. In Chapter 2 an exact experimental design problem is formulated mathematically and is compared with its continuous analogue. Motivation is provided for the solution of this continuous problem, and the remainder of the chapter concerns this problem. A necessary and sufficient condition for optimality of a design measure is given. Problems which might arise in testing this condition are discussed, in particular with respect to possible non-differentiability of the criterion function at the design being tested. Several examples are given of optimal designs which may be found analytically and which illustrate the points discussed earlier in the chapter. In Chapter 3 numerical methods of solution of the continuous optimal design problem are reviewed. A new algorithm is presented with illustrations of how it should be used in practice. It is shown that, for reasonably large sample size, continuously optimal designs may be approximated to well by an exact design. In situations where this is not satisfactory algorithms for improvement of this design are reviewed. Chapter 4 consists of a discussion of sequentially designed experiments, with regard to both the philosophies underlying, and the application of the methods of, statistical inference. In Chapter 5 we criticise constructively previous suggestions for fully sequential design procedures. Alternative suggestions are made along with conjectures as to how these might improve performance. Chapter 6 presents a simulation study, the aim of which is to investigate the conjectures of Chapter 5. The results of this study provide empirical support for these conjectures. In Chapter 7 examples are analysed. These suggest aids to sequential experimentation by means of reduction of the dimension of the design space and the possibility of experimenting semi-sequentially. Further examples are considered which stress the importance of the use of prior information in situations of this type. Finally we consider the design of experiments when semi-sequential experimentation is mandatory because of the necessity of taking batches of observations at the same time. In Chapter 8 we look at some of the assumptions which have been made and indicate what may go wrong where these assumptions no longer hold.

Learning Binary Code Representations for Effective and Efficient Image Retrieval

The size of online image datasets is constantly increasing. Considering an image dataset with millions of images, image retrieval becomes a seemingly intractable problem for exhaustive similarity search algorithms. Hashing methods, which encodes high-dimensional descriptors into compact binary strings, have become very popular because of their high efficiency in search and storage capacity. In the first part, we propose a multimodal retrieval method based on latent feature models. The procedure consists of a nonparametric Bayesian framework for learning underlying semantically meaningful abstract features in a multimodal dataset, a probabilistic retrieval model that allows cross-modal queries and an extension model for relevance feedback. In the second part, we focus on supervised hashing with kernels. We describe a flexible hashing procedure that treats binary codes and pairwise semantic similarity as latent and observed variables, respectively, in a probabilistic model based on Gaussian processes for binary classification. We present a scalable inference algorithm with the sparse pseudo-input Gaussian process (SPGP) model and distributed computing. In the last part, we define an incremental hashing strategy for dynamic databases where new images are added to the databases frequently. The method is based on a two-stage classification framework using binary and multi-class SVMs. The proposed method also enforces balance in binary codes by an imbalance penalty to obtain higher quality binary codes. We learn hash functions by an efficient algorithm where the NP-hard problem of finding optimal binary codes is solved via cyclic coordinate descent and SVMs are trained in a parallelized incremental manner. For modifications like adding images from an unseen class, we propose an incremental procedure for effective and efficient updates to the previous hash functions. Experiments on three large-scale image datasets demonstrate that the incremental strategy is capable of efficiently updating hash functions to the same retrieval performance as hashing from scratch.

PREVALENCE OF IRON DEFICIENCY ANAEMIA AND DIETARY IRON INTAKE AMONG INFANTS AGED SIX TO NINE MONTHS IN KEIYO SOUTH SUB COUNTY, KENYA

The increased prevalence of iron deficiency among infants can be attributed to the consumption of an iron deficient diet or a diet that interferes with iron absorption at the critical time of infancy, among other factors. The gradual shift from breast milk to other foods and liquids is a transition period which greatly contributes to iron deficiency anaemia (IDA). The purpose of this research was to assess iron deficiency anaemia among infants aged six to nine months in Keiyo South Sub County. The specific objectives of this study were: to establish the prevalence of infants with iron deficiency anaemia and dietary iron intake among infants aged 6 to 9 months. The cross sectional study design was adopted in this survey. This study was conducted in three health facilities in Keiyo South Sub County. The infants were selected by use of a two stage cluster sampling procedure. Systematic random sampling was then used to select a total of 244 mothers and their infants. Eighty two (82) infants were selected from Kamwosor sub-district hospital and eighty one (81) from both Nyaru and Chepkorio health facilities. Interview schedules, 24-hour dietary recall and food frequency questionnaires were used for collection of dietary iron intake. Biochemical tests were carried out by use of the Hemo-control photochrometer at the health facilities. Infants whose hemoglobin levels were less than 11g/dl were considered anaemic. Further, peripheral blood smears were conducted to ascertain the type of nutritional anaemia. Data was analyzed using the Statistical Package for Social Sciences (SPSS) computer software version 17, 2009. Dietary iron intake was analyzed using the NutriSurvey 2007 computer software. Results indicated that the mean hemoglobin values were 11.3± 0.84 g/dl. Twenty one percent (21.7%) of the infants had anaemia and further 100% of peripheral blood smears indicated iron deficiency anaemia. Dietary iron intake was a predictor of iron deficiency anaemia in this study (t=-3.138; p=0.01). Iron deficiency anaemia was evident among infants in Keiyo South Sub County. The Ministry of Health should formulate and implement policies on screening for anaemia and ensure intensive nutrition education on iron rich diets during child welfare clinics.

New Rigorous Decomposition Methods for Mixed-integer Linear and Nonlinear Programming

Process systems design, operation and synthesis problems under uncertainty can readily be formulated as two-stage stochastic mixed-integer linear and nonlinear (nonconvex) programming (MILP and MINLP) problems. These problems, with a scenario based formulation, lead to large-scale MILPs/MINLPs that are well structured. The first part of the thesis proposes a new finitely convergent cross decomposition method (CD), where Benders decomposition (BD) and Dantzig-Wolfe decomposition (DWD) are combined in a unified framework to improve the solution of scenario based two-stage stochastic MILPs. This method alternates between DWD iterations and BD iterations, where DWD restricted master problems and BD primal problems yield a sequence of upper bounds, and BD relaxed master problems yield a sequence of lower bounds. A variant of CD, which includes multiple columns per iteration of DW restricted master problem and multiple cuts per iteration of BD relaxed master problem, called multicolumn-multicut CD is then developed to improve solution time. Finally, an extended cross decomposition method (ECD) for solving two-stage stochastic programs with risk constraints is proposed. In this approach, a CD approach at the first level and DWD at a second level is used to solve the original problem to optimality. ECD has a computational advantage over a bilevel decomposition strategy or solving the monolith problem using an MILP solver. The second part of the thesis develops a joint decomposition approach combining Lagrangian decomposition (LD) and generalized Benders decomposition (GBD), to efficiently solve stochastic mixed-integer nonlinear nonconvex programming problems to global optimality, without the need for explicit branch and bound search. In this approach, LD subproblems and GBD subproblems are systematically solved in a single framework. The relaxed master problem obtained from the reformulation of the original problem, is solved only when necessary. A convexification of the relaxed master problem and a domain reduction procedure are integrated into the decomposition framework to improve solution efficiency. Using case studies taken from renewable resource and fossil-fuel based application in process systems engineering, it can be seen that these novel decomposition approaches have significant benefit over classical decomposition methods and state-of-the-art MILP/MINLP global optimization solvers.

Adjustable robust optimization with nonlinear recourses

Over the last century, mathematical optimization has become a prominent tool for decision making. Its systematic application in practical fields such as economics, logistics or defense led to the development of algorithmic methods with ever increasing efficiency. Indeed, for a variety of real-world problems, finding an optimal decision among a set of (implicitly or explicitly) predefined alternatives has become conceivable in reasonable time. In the last decades, however, the research community raised more and more attention to the role of uncertainty in the optimization process. In particular, one may question the notion of optimality, and even feasibility, when studying decision problems with unknown or imprecise input parameters. This concern is even more critical in a world becoming more and more complex —by which we intend, interconnected —where each individual variation inside a system inevitably causes other variations in the system itself. In this dissertation, we study a class of optimization problems which suffer from imprecise input data and feature a two-stage decision process, i.e., where decisions are made in a sequential order —called stages —and where unknown parameters are revealed throughout the stages. The applications of such problems are plethora in practical fields such as, e.g., facility location problems with uncertain demands, transportation problems with uncertain costs or scheduling under uncertain processing times. The uncertainty is dealt with a robust optimization (RO) viewpoint (also known as "worst-case perspective") and we present original contributions to the RO literature on both the theoretical and practical side.