Several studies have been carryout on finding optimal locations to minimize CO2 emissions from the last mile distribution perspective. In conjunction with that, there has been no study conducted in Sweden that provides a decision support system to compute the transport consequences of the modifications in the retailer’s store network. This thesis did used the following steps: requirement analysis, system design, implementation and testing to build a prototype decision support system that is to help retailers find optimal locations for a new retail store. This thesis provided a subsequent answer as to which data are needed along with the rightful user interface for said decision support system. Subsequently, this thesis does present a decision support system prototype from which some recommendations were provided as to what skills set and tools are needed for the management and maintenance of said decision support system. The primary data used during this thesis is the Dalarna municipalities, six selected retailer’s stores networks and the Dalarna Road network geo-data (Longitude and latitude). This thesis does conclude that it is possible to integrate an optimization model within the Django framework using a geo data to build a decision support system.

The Burr families (Type III and Type XII) of distributions are traditionally used in the context of statistical modeling and for simulating non-normal distributions with moment-based parameters (e.g., Skew and Kurtosis). In educational and psychological studies, the Burr families of distributions can be used to simulate extremely asymmetrical and heavy-tailed non-normal distributions. Conventional moment-based estimators (i.e., the mean, variance, skew, and kurtosis) are traditionally used to characterize the distribution of a random variable or in the context of fitting data. However, conventional moment-based estimators can (a) be substantially biased, (b) have high variance, or (c) be influenced by outliers. In view of these concerns, a characterization of the Burr Type III and Type XII distributions through the method of L-moments is introduced. Specifically, systems of equations are derived for determining the shape parameters associated with user specified L-moment ratios (e.g., L-Skew and L-Kurtosis). A procedure is also developed for the purpose of generating non-normal Burr Type III and Type XII distributions with arbitrary L-correlation matrices. Numerical examples are provided to demonstrate that L-moment based Burr distributions are superior to their conventional moment based counterparts in the context of estimation, distribution fitting, and robustness to outliers. Monte Carlo simulation results are provided to demonstrate that L-moment-based estimators are nearly unbiased, have relatively small variance, and are robust in the presence of outliers for any sample size. Simulation results are also provided to show that the methodology used for generating correlated non-normal Burr Type III and Type XII distributions is valid and efficient. Specifically, Monte Carlo simulation results are provided to show that the empirical values of L-correlations among simulated Burr Type III (and Type XII) distributions are in close agreement with the specified L-correlation matrices.

The 911 emergency response process is a core component of the emergency services critical infrastructure sector in the United States. Modeling and simulation of a complex stochastic system like the 911 response process enables policy makers and stakeholders to better understand, identify, and mitigate the impact of attacks/disasters affecting the 911 system. Modeling the 911 response process as a series of queue sub-systems will enable analysis into how CI failures impact the different phases of the 911 response process. Before such a model can be constructed, the probability distributions of the inter-arrivals of events into these various sub-systems needs to be identified. This research is a first effort into investigating the stochastic behavior of inter-arrival times of different events throughout the 911 response process. I use the methodology of input modeling, a statistical modeling approach, to determine whether the exponential distribution is an appropriate model for these inter-arrival times across a large dataset of historical 911 dispatch records.

Biomarker discovery and statistical modeling reveals the brain activity that supports brain function and dysfunction. Detecting abnormal brain activity is critical for developing biomarkers of disease, elucidating disease mechanisms and evolution, and ultimately improving disease course. In my thesis, we develop statistical methodology to characterize neural activity in disease from noisy electrophysiological recordings. First, we develop a modification of a classic statistical modeling approach - multivariate Granger causality - to infer coordinated activity between brain regions. Assuming the signaling dependencies vary smoothly, we propose to write the history terms in autoregressive models of the signals using a lower dimensional spline basis. This procedure requires fewer parameters than the standard approach, thus increasing the statistical power. we show that this procedure accurately estimates brain dynamics in simulations and examples of physiological recordings from a patient with pharmacoresistant epilepsy. This work provides a statistical framework to understand alternations in coordinated brain activity in disease. Second, we demonstrate that sleep spindles, thalamically-driven neural rhythms (9-15 Hz) associated with sleep-dependent learning, are a reliable biomarker for Rolandic epilepsy. Rolandic epilepsy is the most common form of childhood epilepsy and characterized by nocturnal focal epileptic discharges as well as neurocognitive deficits. We show that sleep spindle rate is reduced regionally across cortex and correlated with poor cognitive performance in epilepsy. These results provide evidence for a regional disruption to the thalamocortical circuit in Rolandic epilepsy, and a potential mechanistic explanation for the cognitive deficits observed. Finally, we develop a procedure to utilize delta rhythms (2-4 Hz), a sensitive biomarker for Angelman syndrome, as a non-invasive measure of treatment efficacy in clinical trials. Angelman syndrome is a rare neurodevelopmental disorder caused by reduced expression of the UBE3A protein. Many disease-modifying treatments are being developed to reinstate UBE3A expression. To aid in clinical trials, we propose a procedure that detects therapeutic improvements in delta power outside of the natural variability over age by developing a longitudinal natural history model of delta power. These results demonstrate the utility of biomarker discovery and statistical modeling for elucidating disease course and mechanisms with the long-term goal of improving patient outcomes.

This PhD thesis focuses on current issues of commercial risk in international trade, particularly on the evaluation of commercial risk when exporting to China. This PhD thesis presents initial theoretical framework for solution of the problem and also presents statistical results of primary research conducted for Czech exporters necessary to meet the objectives of the dissertation. The aim of this PhD thesis is to construct the model for assessment of commercial risks of exporting to China. The constructed model is probabilistic model, while outcoming results of resulting commercial risk rating based on the averaging of the probable costs or losses caused by the effects of commercial risk which may arise in exporting business entity at unsecured contract, and may take considerable values. The constructed model allows both, to calculate with costs in their absolute probable values as well as to calculate with costs in their relative values as percentages of the contract value. The issue of trade with China is broad and encompasses several disciplines. This implies a large potential for further research which aims in particular to the modeling of knowledge, and by extension created the probabilistic model to the knowledge model.

Recent improvements in numerical weather model resolution open the possibility of producing forecasts for lightning using indirect lightning threat indicators well in advance of an event. This research examines the feasibility of a statistical machine-learning algorithm known as a support vector machine (SVM) to provide a probabilistic lightning forecast for Mississippi at 9 km resolution up to one day in advance of a thunderstorm event. Although the results indicate that SVM forecasts are not consistently accurate with single-day lightning forecasts, the SVM performs skillfully on a data set consisting of many forecast days. It is plausible that errors by the numerical forecast model are responsible for the poorer performance of the SVM with individual forecasts. More research needs to be conducted into the possibility of using SVM for lightning prediction with input data sets from a variety of numerical weather models.

Cities have the potential to change the way resources and nutrients are utilized as they are centers of consumption and waste production. Losses of nutrients like nitrogen and phosphorus (P) to water ways, called eutrophication, is a major water quality issue that marine ecosystems face (Bennett et al., 2001; Smith & Schindler, 2009). Urban agriculture (UA) provides a chance for some nutrient reuse within city boundaries, but there exists a gap in knowledge regarding how soil properties influence P movement patterns within UA contexts. To explore the relationships between P leachate and soil characteristics from urban gardens, I created generalized linear mixed models (GLMMs) using data from 8 gardens in Linköping, Sweden, over a period of 2 years. Though leachate data and soil traits varied between gardens, values from the urban gardens generally did not vary extensively compared to those from field studies or rural agriculture. As hypothesized, plant-available P from the ammonium lactate soil P test (P-AL) and degree of P saturation (DPS) were both important, although why they were significant to their respective water quality variables was unclear. Moreover, spatial correlations were also not as influential as expected in P leaching. Additionally, other important soil characteristics (pH, clay, plant-available iron (Fe-AL), and plant-available aluminum (Al-AL)) seemed to relate to P adsorption and release, indicating a need for future research in that direction.

This thesis introduces a new effective method in statistical modeling and probabilistic decision making problems. The method is based on maximizing the Shannon Logarithmic Entropy Function for information, subject to the given prior information to serve as constraints, to generate a probability distribution. The method is known as the Maximum Entropy Principle or "Jaynes Principle". Tribus used it earlier, but in a limited case, without general application to either statistical modeling or probablistic decision making. In this thesis, a new method which generalizes the above principle is introduced. This permits practical applications, some of which are illustrated. / Thesis / Master of Engineering (ME)

No description available.

No description available.