Analytical solutions such as the Theis solution have historically been utilized to forecast changes in aquifer water levels resulting from human-driven withdrawals using pumping wells. This method, however, suffers from a number of disadvantages, such as long data acquisition times, model uncertainty, and trial-and-error calibrations. This study illustrated the effectiveness of alternate forecasting methods that utilized machine learning principles. The groundwater level dynamics of two sites located at the Virginia Eastern Shore were predicted using historical groundwater level below land surface (GWLBLS) data as the endogenous variable and local pumping data as the exogenous variable. Predicting the local pumping data from the GWLBLS values was also implemented, to not only enforce reliability of the model, but also to highlight the capability of verifying and enforcing permitted pumping data. The machine learning methods chosen for this study were the Random Forest and SARIMAX models. Historical datasets were divided into training/calibration and testing/validation sets, and the respective models were fit to the data. These calibrated models were then compared to the performance of the Theis solution. Across both study sites, the Random Forest performed best at forecasting groundwater level over time given the pumping data as an exogenous variable, with SARIMAX performing similarly to the Theis solution. The Theis solution, however, did perform well in terms of generalization ability (GA). / Master of Science / Groundwater is a vital resource for drinking, agriculture, and industry. In order to ensure aquifer health for future use, it is crucial to be able to forecast well water level in the midst of groundwater pumping. Currently, analytical solutions such as the Theis solution are utilized to predict water level over time, but data acquisition is time-consuming and many of the calibrations have to be based on trial-and-error. In this thesis, machine learning methods were explored as alternatives to the current analytical methods. The groundwater level dynamics of the two study sites, Oyster, Virginia and Temperanceville, Virginia, were used to calibrate corresponding machine learning models, called the Random Forest (RF) and SARIMAX models. While the Theis showed that it was the most adaptable model, the RF performed the best overall in terms of root mean square error and R2 scores, which were used as reliability metrics. This study provides a range of substitutes for the Theis solution that have the capability to perform better when calibrated on a site-by-site basis.

No description available.

The thesis outlines the use of Orthogonal Arrays for the training of Artificial Neural Networks. Such arrays are popularly used in system optimisation and are known as Taguchi Methods. The chief advantage of the method is that the network can learn quickly. Fast training methods may be used in certain Control Systems and it has been suggested that they could find application in ‘disaster control,’ where a potentially dangerous system (for example, suffering a mechanical failure) needs to be controlled quickly. Previous work on the methods has shown that they suffer problems when used with multi-layer networks. The thesis discusses the reasons for these problems and reports on several successful techniques for overcoming them. These techniques are based on the consideration of the neuron, rather then the individual weight, as a factor to be optimised. The applications of technique and further work are also discussed.

Diabetes Mellitus is a major health problem which affects about 200 million people worldwide. Diabetics require their blood glucose levels to be kept within the normal range in order to prevent diabetes-related complications from occurring. Blood glucose measurement is therefore of vital importance. The current glucose measurement techniques are, however, painful, inconvenient and episodic. This document provides an investigation into the use of near-infrared spectroscopy for continuous, non-invasive measurement of blood glucose. Artificial neural networks are used for the development of multivariate calibration models which predict glucose concentrations based on the near-infrared spectral data. Simulations have been performed which make use of simulated spectral data generated from the characteristic spectra of many of the major components of human blood. The simulations show that artificial neural networks are capable of predicting the glucose concentrations of complex aqueous solutions with clinically relevant accuracy. The effect of interference, such as temperature changes, pathlength variations, measurement noise and absorption due other analytes, has been investigated and modelled. The artificial neural network calibration models are capable of providing acceptably accurate predictions in the presence of multiple forms of interference. It was found that the performance of the measurement technique can be improved through careful selection of the optical pathlength and wavelength range for the spectroscopic measurements, and by using preprocessing techniques to reduce the effect of interference. Although the simulations suggest that near-infrared spectroscopy is a promising method of blood glucose measurement, which could greatly improve the quality of life of diabetics, many further issues must be resolved before the long-term goal of developing a continuous non-invasive home glucose monitor can be achieved.

This study is concerned with the analyses of modern electric power-grids designed to support large supply-demand considerations in metro areas of large cities. Hence proposed are methods to determine optimal performance of the associated distribution networks vis-á-vis power availability from multiple resources (such as hydroelectric, thermal, wind-mill, solar-cell etc.) and varying load-demands posed by distinct set of consumers of domestic, industrial and commercial sectors. Hence, developing the analytics on optimal power-distribution across pertinent power-grids are verified with the models proposed. Forecast algorithms and computational outcomes on supply-demand performance are indicated and illustratively explained using real-world data sets. This study on electric utility takes duly into considerations of both deterministic (technological factors) as well as stochastic variables associated with the available resource-capacity and demand-profile details. Thus, towards forecasting exercise as above, a representative load-curve (RLC) is defined; and, it is optimally determined using an Artificial Neural Network (ANN) method using the data availed on supply-demand characteristics of a practical power-grid. This RLC is subsequently considered as an input parametric profile on tariff policies associated with electric power product-cost. This research further focuses on developing an optimal/suboptimal electric-power distribution scheme across power-grids deployed between multiple resources and different sets of user demands. Again, the optimal/suboptimal decisions are enabled using ANN-based simulations performed on load sharing details. The underlying supply-demand forecasting on distribution service profile is essential to support predictive designs on the amount of power required (or to be generated from single and/or multiple resources) versus distributable shares to different consumers demanding distinct loads. Another topic addressed refers to a business model on a cost reflective tariff levied in an electric power service in terms of the associated hedonic heuristics of customers versus service products offered by the utility operators. This model is based on hedonic considerations and technoeconomic heuristics of incumbent systems In the ANN simulations as above, bootstrapping technique is adopted to generate pseudo-replicates of the available data set and they are used to train the ANN net towards convergence. A traditional, multilayer ANN architecture (implemented with feed-forward and backpropagation techniques) is designed and modified to support a fast convergence algorithm, used for forecasting and in load-sharing computations. Underlying simulations are carried out using case-study details on electric utility gathered from the literature. In all, ANN-based prediction of a representative load-curve to assess power-consumption and tariff details in electrical power systems supporting a smart-grid, analysis of load-sharing and distribution of electric power on smart grids using an ANN and evaluation of electric power system infrastructure in terms of tariff worthiness deduced via hedonic heuristics, constitute the major thematic efforts addressed in this research study. / Includes bibliography. / Dissertation (Ph.D.)--Florida Atlantic University, 2019. / FAU Electronic Theses and Dissertations Collection

Le principal objectif de ce travail est l’étude de l’influence de la pesanteur sur la mécanique respiratoire et le contrôle de la respiration, ainsi que sur les interactions cardio-respiratoires pendant les différents stades du sommeil. Le chapitre introductif présente le contexte général et les objectifs de la thèse. Des sections abordant le sommeil, la respiration, et l’interaction cardio-respiratoire y sont présentées, résumant l’état actuel des connaissances sur les effets de la pesanteur sur chacun de ces systèmes. Dans le deuxième chapitre, l’expérience “Sleep and Breathing in microgravity”, qui constitue la source des données à la base de ce travail, est présentée en détail. L’étude des signaux de longue durée requiert avant tout de disposer d’outils performants d’analyse des signaux. La première partie de la thèse présente en détail deux algorithmes : un algorithme de détection automatique d’événements respiratoires (inspiration / expiration) basé sur des réseaux neuronaux artificiels, et un algorithme de quantification de l’amplitude et de la phase de l’arythmie sinusale pendant le sommeil, utilisant la méthode des ondelettes. La validation de chaque algorithme est présentée, et leur performance évaluée. Cette partie inclut aussi des courtes introductions théoriques aux réseaux de neurones artificiels ainsi qu’aux méthodes d’analyse temps–fréquence (Fourier et ondelettes). Une approche similaire à celle utilisée pour la détection automatique d’événements respiratoires a été appliquée à la détection d’événements dans des signaux de vitesse du sang dans l’artère cérébrale moyenne, mesures obtenues par Doppler transcrânien. Ceci est le sujet de la thèse annexe. Ces deux algorithmes ont été appliqués aux données expérimentales pour extraire des informations physiologiques quant à l’impact de la pesanteur sur la mécanique respiratoire et l’interaction cardio-respiratoire. Ceci constitue la deuxième partie de la thèse. Un chapitre est consacré aux effets de l’apesanteur sur la mécanique respiratoire pendant le sommeil. Ce chapitre a mis en évidence, pour tous les stades de sommeil, une augmentation de la contribution abdominale en microgravité, suivi d’un retour progressif vers des valeurs observées avant le vol. L’augmentation initiale était attendue, mais l’adaptation progressive observée ne peut pas être expliquée par un effet purement mécanique, et nous suggère la présence d’un mécanisme d’adaptation central. Un deuxième chapitre présente les résultats comparant l’arythmie sinusale pendant le sommeil avant le vol, en apesanteur et après le retour sur terre. Le rythme cardiaque pendant le sommeil dans l’espace présente une moindre variabilité. Les différences NREM–REM observées sur terre pour les influences vagales et sympathiques sont accentuées dans l’espace. Aucun changement significatif n’est présent pour le gain et la différence de phase entre les les signaux cardiaque et respiratoire en comparant le sommeil sur terre et en apesanteur. La dissertation termine par une discussion générale du travail effectué, incluant les prin- cipales conclusions ainsi que les perspectives qui en découlent.

Intelligent decision systems have the potential to support and greatly amplify human decision-making across a number of industries and domains. However, despite the rapid improvement in the underlying capabilities of these “intelligent” systems, increasing their acceptance as decision aids in industry has remained a formidable challenge. If intelligent systems are to be successful, and their full impact on decision-making performance realized, a greater understanding of the factors that influence recommendation acceptance from intelligent machines is needed. Through an empirical experiment in the financial services industry, this study investigated the effects of perceived behavioral similarity (similarity state) on the dependent variables of recommendation acceptance, decision performance and decision efficiency under varying conditions of uncertainty (volatility state). It is hypothesized in this study that behavioral similarity as a design element will positively influence the acceptance rate of machine recommendations by human users. The level of uncertainty in the decision context is expected to moderate this relationship. In addition, an increase in recommendation acceptance should positively influence both decision performance and decision efficiency. The quantitative exploration of behavioral similarity as a design element revealed a number of key findings. Most importantly, behavioral similarity was found to positively influence the acceptance rate of machine recommendations. However, uncertainty did not moderate the level of recommendation acceptance as expected. The experiment also revealed that behavioral similarity positively influenced decision performance during periods of elevated uncertainty. This relationship was moderated based on the level of uncertainty in the decision context. The investigation of decision efficiency also revealed a statistically significant result. However, the results for decision efficiency were in the opposite direction of the hypothesized relationship. Interestingly, decisions made with the behaviorally similar decision aid were less efficient, based on length of time to make a decision, compared to decisions made with the low-similarity decision aid. The results of decision efficiency were stable across both levels of uncertainty in the decision context.

Intelligent decision systems have the potential to support and greatly amplify human decision-making across a number of industries and domains. However, despite the rapid improvement in the underlying capabilities of these “intelligent” systems, increasing their acceptance as decision aids in industry has remained a formidable challenge. If intelligent systems are to be successful, and their full impact on decision-making performance realized, a greater understanding of the factors that influence recommendation acceptance from intelligent machines is needed. Through an empirical experiment in the financial services industry, this study investigated the effects of perceived behavioral similarity (similarity state) on the dependent variables of recommendation acceptance, decision performance and decision efficiency under varying conditions of uncertainty (volatility state). It is hypothesized in this study that behavioral similarity as a design element will positively influence the acceptance rate of machine recommendations by human users. The level of uncertainty in the decision context is expected to moderate this relationship. In addition, an increase in recommendation acceptance should positively influence both decision performance and decision efficiency. The quantitative exploration of behavioral similarity as a design element revealed a number of key findings. Most importantly, behavioral similarity was found to positively influence the acceptance rate of machine recommendations. However, uncertainty did not moderate the level of recommendation acceptance as expected. The experiment also revealed that behavioral similarity positively influenced decision performance during periods of elevated uncertainty. This relationship was moderated based on the level of uncertainty in the decision context. The investigation of decision efficiency also revealed a statistically significant result. However, the results for decision efficiency were in the opposite direction of the hypothesized relationship. Interestingly, decisions made with the behaviorally similar decision aid were less efficient, based on length of time to make a decision, compared to decisions made with the low-similarity decision aid. The results of decision efficiency were stable across both levels of uncertainty in the decision context.

The research presented in this thesis is concerned with optimising the structure of Artificial Neural Networks. These techniques are based on computer modelling of biological evolution or foetal development. They are known as Evolutionary, Genetic or Embryological methods. Specifically, Embryological techniques are used to grow Artificial Neural Network topologies. The Embryological Algorithm is an alternative to the popular Genetic Algorithm, which is widely used to achieve similar results. The algorithm grows in the sense that the network structure is added to incrementally and thus changes from a simple form to a more complex form. This is unlike the Genetic Algorithm, which causes the structure of the network to evolve in an unstructured or random way. The thesis outlines the following original work: The operation of the Embryological Algorithm is described and compared with the Genetic Algorithm. The results of an exhaustive literature search in the subject area are reported. The growth strategies which may be used to evolve Artificial Neural Network structure are listed. These growth strategies are integrated into an algorithm for network growth. Experimental results obtained from using such a system are described and there is a discussion of the applications of the approach. Consideration is given of the advantages and disadvantages of this technique and suggestions are made for future work in the area. A new learning algorithm based on Taguchi methods is also described. The report concludes that the method of incremental growth is a useful and powerful technique for defining neural network structures and is more efficient than its alternatives. Recommendations are also made with regard to the types of network to which this approach is best suited. Finally, the report contains a discussion of two important aspects of Genetic or Evolutionary techniques related to the above. These are Modular networks (and their synthesis) and the functionality of the network itself.

This thesis describes a novel approach to the evolution of Modular Artificial Neural Networks. Standard Evolutionary Algorithms, used in this application include: Genetic Algorithms, Evolutionary Strategies, Evolutionary Programming and Genetic Programming; however, these often fail in the evolution of complex systems, particularly when such systems involve multi-domain sensory information which interacts in complex ways with system outputs. The aim in this work is to produce an evolutionary method that allows the structure of the network to evolve from simple to complex as it interacts with a dynamic environment. This new algorithm is therefore based on Incremental Evolution. A simulated model of a legged robot was used as a test-bed for the approach. The algorithm starts with a simple robotic body plan. This then grows incrementally in complexity along with its controlling neural network and the environment it reacts with. The network grows by adding modules to its structure - so the technique may also be termed a Growth Algorithm. Experiments are presented showing the successful evolution of multi-legged gaits and a simple vision system. These are then integrated together to form a complete robotic system. The possibility of the evolution of complex systems is one advantage of the algorithm and it is argued that it represents a possible path towards more advanced artificial intelligence. Applications in Electronics, Computer Science, Mechanical Engineering and Aerospace are also discussed.