Transportation, Logistics, & Finance Doctoral Work

Permanent URI for this collectionhdl:10365/32389

Browse

Now showing 1 - 8 of 8

Contributing Factors of DUI Recidivism Among First-Time Offenders in North Dakota
(North Dakota State University, 2022) Zhou, Yun
This study explored utilizing tree-based machine learning models to identify associations in a range of 107 factors and DUI recidivism among first-time DUI offenders. Three tree-based machine learning models, Decision Tree, Random Forest, and Gradient Boosting were performed on 12,879 first-time DUI offenders during 2013-2017 using a three-year following period, to classify repeat DUI offenders. Study cohorts include 11,651 drivers without recidivism and 1,228 drivers with recidivism. The models tested 107 variables/predictors, including the driver’s demographic factors, drinking behaviors, traffic violations, crash histories, DUI-related violations, social-economic factors, and health and safety factors based on the driver’s residence. oversampling technique was used to balance two classes in the training data in all three models. The top 15-20 predictors were selected from the feature impact analyses of these predictions. Lastly, multiple logistic regression analyses were performed to quantify the effects of selected factors/predictors on the outcome. Among the three models, Gradient Boosting achieved the best predictions on both the original and oversampled datasets. Oversample techniques did improve prediction performances by roughly 10% on the F1 score for Gradient Boosting. Results coalesced around two findings. First, male drivers with higher BAC values, younger age at first DUI citation, whose first DUI citation took place during the weekday, had at least one low-risk citation within three years before first DUI citation, and lived in counties with lower income inequality ratio and higher violent crime rate were more likely to commit a subsequent DUI offense. Second, male drivers who complied with a BAC test upon arrest, whose first DUI citation took place on a weekday, had at least one low-risk citation within three years before the first DUI citation, lived in a county with a lower income inequality ratio, and higher violent crime rate were more likely to commit a subsequent DUI offense. Findings can be used by stakeholders in implementing and improving DUI prevention strategies. The study is limited to a single state, but the comparison of techniques and their shared findings suggest that a multitude and variety of approaches may be appropriate in future impaired driving prevention research.
Evaluating the Performance of Emergency Medical System in the US
(North Dakota State University, 2022) Ebrahimi, Zhila Dehdari
New and exciting opportunities are emerging for operational researchers to create and use models that provide managers with solutions to enhance the quality of their services as the importance of the service sector grows in industrialized countries. The key to this process is the creation of time-dependent models that analyze complicated service systems and produce efficient staff schedules, allowing organizations to strike a balance between delivering high-quality services and avoiding unnecessary personnel costs. There is a need, particularly in the healthcare sector, to encourage effective management of an EMS, where the likelihood of survival is strongly correlated with the response time.Motivated by case studies investigating the operation of the Emergency Medical System (EMS), this dissertation aims to examine how operations research (OR) techniques can be developed to determine staff scheduling and maximize the ambulance to decrease service system delays. A capacity planning tool is developed that integrates a combination of queueing theory and optimization techniques to reduce the delay in the service system and maximize ambulance coverage. The research presented in this dissertation is novel in several ways. Primarily, the first section considers the Markovian models with sinusoidal arrival rates and state-dependency of service rate and uses a numerical method known as Stationary Independent Period by Period (SIPP) to determine the staff requirement of the service system. The final section considers the time dependency in locating an ambulance station across the network and allocating the ambulance to the patients to cover more 911 calls.
Logistics of North Dakota State Local Food System: An Empirical Study to Measure Local Food Supply and Demand for Regional Food Hub Feasibility Study
(North Dakota State University, 2021) AlQublan, Hamad
Because North Dakota (ND) is one of only four states that do not have a food hub, the purpose of the regional food hub feasibility study was to empirically evaluate whether the ND local food system needs a food hub from a supply and demand perspective. The 21st century has shifted American agriculture and rural life. The number of small and mid-sized farms decreases while the number of more significant farms increases. The food hubs concept has become an increasingly popular response to this agricultural problem. Food hubs cover the gap between farmers and markets and add value to the food supply chain infrastructure. The food hub concept is widely distributed among US states as a response to solving local food issues. Unfortunately, ND is one of the four states that do not have a food hub. Hence, this study's main exploratory research question was Does the ND food system need a regional food hub? Furthermore, the author of this research found no comprehensive literature review concentrating solely on the local food system in ND. For that, a regional food hub feasibility study was conducted. The ND regional food hub feasibility was divided into two independent cross-sectional surveys. The ND food hub feasibility was divided part-A (the supply-side) and part-B (the demand-side). Each survey had 51 questions, including qualitative and quantitative factors. Investigating both sides of the ND local food system led to a better comprehension of the research area and provided a complete understanding of the ND food system. Both surveys were statistically analyzed descriptively and inferentially. Our findings indicated that ND food producers and customers defined local food as all food produced or grown in ND. Additionally, we found that the regional ND food hub project was feasible. The ND regional food hub's suggested best model had a cooperative legal structure and a hybrid business structure that can work for-profit at both the state and national levels.
Optimal Resource Allocation to Minimize Last Mile Delivery Costs
(North Dakota State University, 2022) Srinivasan, Raghavan
This study focusses on a decision-making tool to assist an organization in planning for capacity needed for the Last Mile Delivery (LMD) services which is the most expensive part of the entire supply chain. Considering the use of Crowdsourcing for Logistics (CSL), the decision-making tool’s objective is to provide an optimal combination of fulltime, seasonal and CSL resources that lead to minimum operational LMD costs and meet the variable demand. To achieve this, a three phased approach is used, where in the first analytical phase an expected cost model is numerically validated. In the second stochastic program phase, the capacity and cost of the CSL resources are varied. Finally, in the third simulation phase, the approach is further extended to consider the daily employee attrition rate and unsatisfied demand being carried over to the next day. Lastly, the use of automation or newer technologies, such as robots, for LMD services is introduced in this simulation phase to show the benefits in terms of the operations costs. The results from the analytical model described the optimal values of fulltime and seasonal considering the utilization of CSL and experienced some penalty costs. In this case, the parameters being fixed, does not capture the differences due to the variability of CSL availability or costs, which is addressed in the stochastic program phase. Though the output from the stochastic model is higher, it does consider the variability in the CSL capacity and cost, which is practically observed. The simulation section gives a further refined optimal combination of fulltime, seasonal and CSL that meets the demand considering the attrition rate of fulltime and seasonal, and rollover the units by one day. Within this simulation, the consideration of automated delivery systems like using a robot for LMD services leads to further cost savings opportunity. Here, the fulltime delivery cost is benefited, with low utilization of seasonal and CSL limited for optimizing delivery strategy. In conclusion a tool is provided for aggregate delivery capacity planning that would consider an optimal combination of fulltime, seasonal and CSL resources lowering the LMD costs and meeting the variable demand.
Safety Management System for Highway-Rail Grade Crossings
(North Dakota State University, 2021) Keramati, Amin
As a result of the considerable differences in mass between vehicles and trains, accidences at highway-rail grade crossings (HRGCs) may result in severe injuries and fatalities. Therefore, HRGCs safety is considered one of the crucial transportation safety issues. Transportation decision makers and agencies need an efficient safety decision-making framework which is bale to consider crash occurrence and severity likelihood simultaneously. This study proposed a novel methodology and a statistical approach for HRGC crash analysis. The proposed method is competing risk model and the approach is Cox proportional hazard regression. This predictive method was well established in bioscience area but never been utilized in transportation area. Competing Risk Model (CRM) is a special type of survival analysis to accommodate the competing nature of multiple outcomes from the same event of interest, in transportation safety analysis the competing multiple outcomes are accident severity levels while the event of interest is accident occurrence. Transportation decision makers need a prioritization system to categorize crossings’ risk level based on their expected crash frequency and crash severity simultaneously. Therefore, with a hazard-ranking approach which considers crossings’ crash severity and frequency output, transportation decision makers are able to ensure that federal and state funds for grade crossing improvement projects are spent at the crossings that are considered the most in need of improvement. In this study two hazard-ranking models are proposed, the first one is based on the crash likelihood resulted by the proposed CRM output, and the second one is a hybrid accident prediction model/hazard index based on crash severity likelihoods estimated by the same CRM. Finally, to integrate the results of both hazard-ranking approaches, and classify grade crossings and crossings’ location based on their crash frequency and severity likelihood simultaneously, the risk analysis is conducted by using the risk matrix and spatial risk analysis.
Three Essays on Railroad Safety Analysis Using Non-Parametric Statistical Methods
(North Dakota State University, 2022) Dhingra, Neeraj
he FRA mandated railroad companies to install a new monitoring system known as Positive Train Control (PTC). This system overlays sensors, signals, and transponders over existing track and other wayside infrastructure. Technologists designed the system to prevent accidents mainly caused by human negligence and communications. However, PTC will not address track-related defects, which is the second dominant cause of accidents. A new track monitoring system called Railway Autonomous Inspection Localization System (RAILS) was proposed to address track-related accidents. RAILS is based on low-cost sensor technology that identifies defect symptoms, ranks their severity, classifies defect types, and localizes their positions. So, RAILS technology can augment the PTC by identifying track-related issues. The main objectives of this dissertation are: (1) To compare the potential performance of RAILS with traditional inspection methods based on its fundamental theory of operation; (2) To identify factors contributing to railroad accidents; and (3) To determine and rank factors responsible for severe financial damages caused by railroad accidents.The first two objectives will help compare the proposed technology and identify the major factors responsible for causing train accidents. The final objective will help to categorize accidents based on the potential financial damage severity. Categorizing such incidents would help to create a database that prioritizes issues and suggest possible countermeasure based on the problems. The study's key findings are as follows: (1) RAILS is more efficient in conducting continuous inspection and identifying potential defects than traditional systems by 33%, with only two trains per day and a 50% first-pass detection probability; (2) Nonparametric methods provide implicit information about rail accidents and function better than parametric methods by highlighting factors that are responsible for causing accidents rather than identifying the cause-and-effect relationship; (3) The most significant reasons for causing the financial damages are the number of derailed freight cars and the absence of territory signalization; and (4) Nonparametric methods automatically categorize rail accidents and, using text narratives, highlight causative factors responsible for a train derailment.
Three Essays on Renewable Jet Fuel Supply Chain Network Design and Traffic Safety
(North Dakota State University, 2022) Ebrahimi, Sajad
Because of rising energy consumption, climate change, and environmental concerns about fossil fuels, finding alternative renewable energy sources is becoming increasingly crucial. With the non-advanced share of the U.S. Renewable Fuel Standard having been mainly met by corn ethanol, many states are considering cellulosic or non-edible oilseed crops as the next source of biofuels. This study seeks to design a supply chain to produce renewable jet fuel (RJF) within the Midwest region and southeastern U.S. This is accomplished through the use of optimization models (mixed-integer linear programming). Furthermore, because RJF manufacturing incurs higher expenses than conventional jet fuel, the use of various monetary incentives is being studied to establish their usefulness in commercializing the supply chain. The findings of this study can be used by energy policymakers, RJF producers, and investors to operate in a competitive market while safeguarding the environment. In another study, we evaluate speeding crash risk in North Dakota counties. In the United States, one of the most common contributing factors to car crashes is speeding. Speeding impairs a driver's ability to control and steer properly, as well as respond to a dangerous situation in a timely manner. Speeding crashes account for one-third of fatal crashes in the United States and are one of the risks for drivers on U.S. highways. Speeding crash risk can vary among regions. When it comes to allocating road safety expenditures to regions in order to reduce speeding crashes, it's vital for road management to understand which areas are at higher risk and should be prioritized for safety measures. This study uses a failure mode effect analysis method to evaluate the speeding crash risk.
Use Agreements and Economic Performance of U.S. Airports
(North Dakota State University, 2020) Karanki, Fecri
In the U.S., airport use agreements are developed based on three common rate-setting approaches: the residual, compensatory, and hybrid methods. Under a residual agreement, the financial risk of the host airport is borne by the signatory airlines, and in return, the signatory airlines pay reduced user fees. Under a compensatory agreement, however, the airport bears its own financial risks and offers no reduced user fees to airlines. A hybrid agreement combines the features of residual and compensatory agreements. Under a hybrid agreement, the airport usually bears its own financial risks in terminal operations while the signatory airlines take over the financial risks in airfield operations. This dissertation aims to contribute to air transportation literature concerning the implication of use agreements on airport economic performance and rate differentials. Using the data of 59 U.S. hub airports from years 2009 to 2016, I studied the effects of use agreements on airport operational efficiency (in Chapter 2) and on cost efficiency (in Chapter 3), as well as the sources of aeronautical charge differentials between use agreements (in Chapter 4). The major findings of this dissertation are (1) airports with residual-type agreements tend to have lower operational efficiency compared to their peers adopting either the compensatory or hybrid agreement; (2) airports adopting the residual rate-setting method is less cost-efficient than the airports adopting either the hybrid or compensatory method; (3) compensatory airports have the highest average aeronautical and non-aeronautical charges; (4) non-aeronautical charges are a significant determinant of compensatory airports’ aeronautical charges; (5) airports adopting the hybrid method have lower aeronautical charges than the airports adopting the other two methods due to differences in the average cost level. The first two results imply that under a residual agreement, increased airport inefficiency may undercut any potential benefits of signatory airlines, and this result may indicate the presence of a moral hazard problem in the contractual relationship between the airlines and airports as a result of unequal risk-sharing and information symmetry.

Browse

Browsing Transportation, Logistics, & Finance Doctoral Work by browse.metadata.program "Transportation"