The engineering design community recognizes that an essential part of the design process is decision making. Because decisions are generally made under uncertainty, engineers need appropriate methods for modeling and managing uncertainty. Two important characteristics of uncertainty in the context of engineering design are imprecision and irreducible uncertainty. In order to model both of these characteristics, it is valuable to use probabilities that are most generally imprecise and subjective. These imprecise probabilities generalize traditional, precise probabilities; when the available information is extensive, imprecise probabilities reduce to precise probabilities. An approach for comparing the practical value of different uncertainty models is developed. The approach examines the value of a model using the principles of information economics: value equals benefits minus costs. The benefits of a model are measured in terms of the quality of the product that results from the design process. Costs are measured not only in terms of direct design costs, but also the costs of creating and using the model. Using this approach, the practical value of using an uncertainty model that explicitly recognizes both imprecision and irreducible uncertainty is demonstrated in the context of a high-risk engineering design example in which the decision-maker has few statistical samples to support the decision. It is also shown that a particular imprecise probability model called probability bounds analysis generalizes sensitivity analysis, a process of identifying whether a particular decision is robust given the decision makers lack of complete information. An approach for bounding the value of future statistical data samples while collecting information to support design decisions is developed, and specific policies for making decisions in the presence of imprecise information are examined in the context of engineering.

When modeling a neuron, modelers often focus on the values of parameters that produce a desired output. However, if these parameters are not unique, there could be a number of parameter sets that produce the same output. Thus, even though the values of the various maximum conductances, half activation voltages and so on differ, as a set they can produce the same spike height, firing rates, and so forth. To examine whether or not parameter sets are unique, a 3-compartment motoneuron model was created that has 15 target outputs and 59 parameters. Using parameter searches, over one hundred parameter sets were created for this model that produced the same output (within tolerances). Parameter values vary between parameter sets and indicate that the parameter values are not unique. In addition, some parameters are more tightly constrained than others. Principal component analysis is used to examine the dimensionality of the input and output spaces. However, neurons are more than static output generators. For example, a variety of neuromodulatory influences are known to shift parameter values to alter neuronal output. Thus the question arises as to whether this non-uniqueness extends from model outputs to the models sensitivities to its parameters. In this work, the non-unique parameter sets are further analyzed using sensitivity analyses and output correlations to show that these values vary significantly between these parameter sets. Therefore, each of these models will react to parameter variation differently. This work concludes that parameter sets are non-unique but have varying sensitivity analyses and output correlations. The ramifications of this are discussed for both modelers and neuroscientists.

This thesis is based on Gordon-Schaefer model, and assesses Argentine shortfin squid¡¦s stock by using the data of Southwest Atlantic from FAO between 1983 and 2009. First, estimate the equilibrium level of the open-access fishery and dynamic optimization fishery and compare to each other. Then estimate annual Argentine shortfin squid¡¦s stock size, comparing the stock size with the equilibrium level of the two fishery models. The result is that Argentine shortfin squid¡¦s stock size has no crisis of extinction now in Southwest Atlantic. In addition, simulate Argentine shortfin squid¡¦s stock size under management and no management status in the future. The result is that it will make the Argentine shortfin squid sustainable development under dynamic optimization fishery, and this fishery model will be a good management. Finally, this thesis based on the catch of Southwest Atlantic Argentine shortfin squid, which we figure out the fluctuation of catch by literatures, and do the sensitivity analysis.

Complex engineering systems involve large numbers of functional elements. Each functional element can exhibit complex behavior itself. Ensuring the ability of such systems to meet the customer's needs and requirements requires modeling the behavior of these systems. Behavioral modeling allows a quantitative assessment of the ability of a system to meet specific requirements. However, modeling the behavior of complex systems is difficult due to the complexity of the elements involved and more importantly the complexity of these elements' interactions. In prior work, formal functional modeling techniques have been applied as a means of performing a qualitative decomposition of systems to ensure that needs and requirements are addressed by the functional elements of the system. Extending this functional decomposition to a quantitative representation of the behavior of a system represents a significant opportunity to improve the design process of complex systems. To this end, a functionality-based behavioral modeling framework is proposed along with a sensitivity analysis method to support the design process of complex systems. These design tools have been implemented in a computational framework and have been used to model the behavior of various engineering systems to demonstrate their maturity, application and effectiveness. The most significant result is a multi-fidelity model of a hybrid internal combustion-electric racecar powertrain that enabled a comprehensive quantitative study of longitudinal vehicle performance during various stages in the design process. This model was developed using the functionality-based framework and allowed a thorough exploration of the design space at various levels of fidelity. The functionality-based sensitivity analysis implemented along with the behavioral modeling approach provides measures similar to a variance-based approach with a computation burden of a local approach. The use of a functional decomposition in both the behavioral modeling and sensitivity analysis significantly contributes to the flexibility of the models and their application in current and future design efforts. This contribution was demonstrated in the application of the model to the 2009 Texas A&M Formula Hybrid powertrain design.

Several methods have been developed previously for estimating cumulative energy production and plutonium production from graphite-moderated reactors. The Graphite Isotope Ratio Method (GIRM) is one well-known technique. This method is based on the measurement of trace isotopes in the reactor’s graphite matrix to determine the change in their isotopic ratios due to burnup. These measurements are then coupled with reactor calculations to determine the total plutonium and energy production of the reactor. To facilitate sensitivity analysis of these methods, a one-group cross section and fission product yield library for the fuel and graphite activation products has been developed for MAGNOX-style reactors. This library is intended for use in the ORIGEN computer code, which calculates the buildup, decay, and processing of radioactive materials. The library was developed using a fuel cell model in Monteburns. This model consisted of a single fuel rod including natural uranium metal fuel, magnesium cladding, carbon dioxide coolant, and Grade A United Kingdom (UK) graphite. Using this library a complete sensitivity analysis can be performed for GIRM and other techniques. The sensitivity analysis conducted in this study assessed various input parameters including 235U and 238U cross section values, aluminum alloy concentration in the fuel, and initial concentrations of trace elements in the graphite moderator. The results of the analysis yield insight into the GIRM method and the isotopic ratios the method uses as well as the level of uncertainty that may be found in the system results.

In recent years, Artificial Neural Networks (ANN) have become widely popular tools in various disciplines of engineering, including civil engineering. In this thesis, Multi-layer perceptron with back-propagation type of network is utilized in calibration of the finite element model of a long span cantilever through truss called Commodore Barry Bridge (CBB). The essence of calibration lies in the phenomena of comparing and correlating the structural response of an analytical model with experimental results as closely as possible. Since CBB is a very large structure having complex structural mechanisms, formulation of mathematical expressions representing the relation between dynamics of the structure and the structural parameters is very complicated. Furthermore, when the errors in the structural model and noise in the experimental data are taken into account, a calibration study becomes more tedious. At this point, ANNs are useful tools since they have the capability of learning with noisy data and ability to approximate functions. In this study, firstly sensitivity analyses are conducted such that variations in dynamic properties of the bridge are observed with the changes in its structural parameters. In the second part, inverse relation between sensitive structural parameters and modal frequencies of CBB is approximated by training of a neural network. This successfully trained network is then fed up with experimental frequencies to acquire the as-is structural parameters and model updating is achieved accordingly.

Vehicle handling models are commonly used in the design and analysis of vehicle dynamics. Especially, with the advances in vehicle control systems need for accurate and simple vehicle handling models have increased. These models have parameters, some of which are known or easily obtainable, yet some of which are unknown or difficult to obtain. These parameters are obtained by system identification, which is the study of building model from experimental data. In this thesis, identification of vehicle handling models is based on data obtained from the simulation of complex vehicle dynamics model from ADAMS representing the real vehicle and a general methodology has been developed. Identified vehicle handling models are the linear bicycle model and vehicle roll models with different tire models. Changes of sensitivity of the model outputs to model parameters with steering input frequency have been examined by sensitivity analysis to design the test input. To show that unknown parameters of the model can be identified uniquely, structural identifiability analysis has been performed. Minimizing the difference between the data obtained from the simulation of ADAMS vehicle model and the data obtained from the simulation of simple handling models by mathematical optimization methods, unknown parameters have been estimated and handling models have been identified. Estimation task has been performed using MATLAB Simulink Parameter Estimation Toolbox. By model validation it has been shown that identified handling models represent the vehicle system successfully.

Distortion (undesired dimension and shape changes) is one of the most important problems of through hardened steel components. During quenching, anisotropic dimensional changes are inevitable due to classical plasticity and transformation induced plasticity. Moreover / various distortion potential carriers are brought into material during production chain. This study consists of analytical, numerical and experimental investigations of quench distortion. In numerical and analytical part, sensitivity analysis of the quenching model, and dimensional analysis of distortion were conducted by utilizing experimentally verified simulations. In sensitivity analysis, effect of uncertainties in input data on simulation results were determined, whereas / in dimensional analysis, the influence of various dimensionless numbers that govern quench distortion were investigated. Throughout the study, gas-nozzle-field quenching of SAE52100 long shafts were simulated. Simulations were performed by commercial finite element analysis software, SYSWELD&reg / . Conceptual results indicate that the most important material properties and dimensionless numbers are the ones that govern volume change. Moreover, those that determine plasticity of austenite significantly affect isotropy of the dimensional changes. When unimportant dimensionless numbers are eliminated, there remain 14 dimensionless combinations that govern the problem. In experimental part of the study / effect of microstructure on distortion behavior of SAE52100 long cylinders with various diameters was investigated. In addition to gas-nozzle-field quenching, salt bath and high speed quenching experiments were performed. In regards to experimental findings, there is a correlation between distortions of long cylinders and machining position with respect to billet.

An ecosystem always involves some implicit relations between habitat environment and inhabitants, whose reciprocal links can not be identified easily. Three sets of ecological monitoring data were analyzed in this study, including coral reef, algae (Thalassia hemprichii Aschers) in Kenting National Park, and Formosan landlocked salmon (Oncorhynchus masou formosanus) in the basin of Chichiawan Stream. Two data-oriented analysis techniques, which are Habitat Evaluation Procedure (HEP) and Group Method of Data Handling (GMDH), were applied to retrieve the embedded patterns from these data sets. Eventually, for each data set, a forecasting model based on the technique of combined forecasting were developed, which is to integrate the results from HEP and GMDH, for improving the overall modeling precision. The results of this study show that the data-oriented analyses, such as HEP and GMDH, are useful for finding valid information from the ecological data. Furthermore, the combined forecasting technique can really improve the performance of model prediction even for the ecological research. In order to acquire the most important habitat environmental factors affecting the inhabitants, this study also performed sensitivity analysis of the models. The contributions of this study are to identify effective knowledge for future ecological research and to provide reasonable suggestions for formulating conservation strategy.

<p>This thesis concerns optimization of structures considering various uncertainties. The overall objective is to find methods to create solutions that are optimal both in the sense of handling the typical load case and minimising the variability of the response, i.e. robust optimal designs.</p><p>Traditionally optimized structures may show a tendency of being sensitive to small perturbations in the design or loading conditions, which of course are inevitable. To create robust designs, it is necessary to account for all conceivable variations (or at least the influencing ones) in the design process.</p><p>The thesis is divided in two parts. The first part serves as a theoretical background to the second part, the two appended articles. This first part includes the concept of robust design, basic statistics, optimization theory and meta modelling.</p><p>The first appended paper is an application of existing methods on a large industrial example problem. A sensitivity analysis is performed on a Scania truck cab subjected to impact loading in order to identify the most influencing variables on the crash responses.</p><p>The second paper presents a new method that may be used in robust optimizations, that is, optimizations that account for variations and uncertainties. The method is demonstrated on both an analytical example and a Finite Element example of an aluminium extrusion subjected to axial crushing.</p> / ROBDES