Recurring incidents between pedestrians, bicycles, and vehicles at the intersection of Rural Road and Spence Avenue led to a team of students conducting their own investigation into the current conditions and analyzing a handful of alternatives. An extension of an industry-standard technique was used to build a control case which alternatives would be compared to. Four alternatives were identified, and the two that could be modeled in simulation software were both found to be technically feasible in the preliminary analysis.

ABSTRACT

This study examines the methodology for converting protected, permissive, and protected/permissive left-turn operation to flashing yellow arrow left-turn operation. This study addresses construction-related considerations, including negative offsets, lateral traffic signal head position, left-turn accident rates, crash modification factors and crash reductions factors. A total of 85 intersections in Glendale, Arizona were chosen for this study. These intersections included 45 “arterial to arterial” intersections (a major road intersecting with a major road) and 40 “arterial to collector” intersections (a major road intersecting with a minor road).

This thesis is a clinical study of the field conversion to flashing yellow arrow traffic signals and is not a study of the merits of flashing yellow arrow operation. This study included six categories: 1. High accident intersections (for inclusion in Highway Safety Improvement Program (HSIP) funding); 2. Signal head modifications only; 3. Signal head replacement with median modifications; 4. Signal head and mast arm replacement; 5. Signal head, signal pole and mast arm replacement; and 6. Intersections where flashing yellow arrow operation is not recommended. Compliance with the Manual on Uniform Traffic Control Devices (MUTCD) played a large part in determining conversion costs because the standard for lateral position of the left-turn traffic signal greatly influenced the construction effort. Additionally, the left-turning vehicle’s sight distance factored into cost considerations. It’s important for agencies to utilize this study to understand all of the financial commitments and construction requirements for conversion to flashing yellow arrow operation, and ultimately to appreciate that the process is not purely a matter of swapping traffic signal heads.

Given that more and more planned special events are hosted in urban areas, during which travel demand is considerably higher than usual, it is one of the most effective strategies opening public rapid transit lines and building park-and-ride facilities to allow visitors to park their cars and take buses to the event sites. In the meantime, special event workforce often needs to make balances among the limitations of construction budget, land use and targeted travel time budgets for visitors. As such, optimizing the park-and-ride locations and capacities is critical in this process of transportation management during planned special event. It is also known as park-and-ride facility design problem.

This thesis formulates and solves the park-and-ride facility design problem for special events based on space-time network models. The general network design process with park-and-ride facilities location design is first elaborated and then mathematical programming formulation is established for special events. Meanwhile with the purpose of relax some certain hard constraints in this problem, a transformed network model which the hard park-and-ride constraints are pre-built into the new network is constructed and solved with the similar solution algorithm. In doing so, the number of hard constraints and level of complexity of the studied problem can be considerable reduced in some cases. Through two case studies, it is proven that the proposed formulation and solution algorithms can provide effective decision supports in selecting the locations and capabilities of park-and-ride facilities for special events.

Managed Lanes (MLs) have been increasingly advocated as a way to reduce congestion. This study provides an innovative new tolling strategy for MLs called the travel time refund (TTR). The TTR is an “insurance” that ensures the ML user will arrive to their destination within a specified travel time savings, at an additional fee to the toll. If the user fails to arrive to their destination, the user is refunded the toll amount.

To gauge interest in the TTR, a stated preference survey was developed and distributed throughout the Phoenix-metropolitan area. Over 2,200 responses were gathered with about 805 being completed. Exploratory data analysis of the data included a descriptive analysis regarding individual and household demographic variables, HOV usage and satisfaction levels, HOT usage and interests, and TTR interests. Cross-tabulation analysis is further conducted to examine trends and correlations between variables, if any.

Because most survey takers were in Arizona, the majority (53%) of respondents were unfamiliar with HOT lanes and their practices. This may have had an impact on the interest in the TTR, although it was not apparent when looking at the cross-tabulation between HOT knowledge and TTR interest. The concept of the HOT lane and “paying to travel” itself may have turned people away from the TTR option. Therefore, similar surveys implementing new HOT pricing strategies should be deployed where current HOT practices are already in existence. Moreover, introducing the TTR concept to current HOT users may also receive valuable feedback in its future deployment.

Further analysis will include the weighting of data to account for sample bias, an exploration of the stated preference scenarios to determine what factors were significant in peoples’ choices, and a predictive model of those choices based on demographic information.

In order to address concerns about the dominance of petroleum-fueled vehicles, the transition to alternative-fueled counterparts is urgently needed. Top barriers preventing the widespread diffusion of alternative-fuel vehicles (AFV) are the limited range and the scarcity of refueling or recharging infrastructures in convenient locations. Researchers have been developing models for optimally locating refueling facilities for range-limited vehicles, and recently a strategy has emerged to cluster refueling stations to encourage consumers to purchase alternative-fuel vehicles by building a critical mass of stations. However, clustering approaches have not yet been developed based on flow-based demand. This study proposes a Threshold Coverage extension to the original Flow Refueling Location Model (FRLM). The new model optimally locates p refueling stations on a network so as to maximize the weighted number of origin zones whose refuelable outbound round trips exceed a given threshold, thus to build critical mass based on flow-based demand on the network. Unlike other clustering approaches, this model can explicitly ensure that flow demands “covered” in the model are refuelable considering the limited driving range of AFVs. Despite not explicitly including local intra-zonal trips, numerical experiments on a statewide highway network proved the effectiveness of the model in clustering stations based on inter-city flow volumes on the network. The model’s policy implementation will provide managerial insights for some key concerns of the industry, such as geographic equity vs. critical mass, from a new perspective. This project will serve as a step to support a more successful public transition to alternative-fuel vehicles.

This dissertation research contributes to the advancement of activity-based travel forecasting models along two lines of inquiry. First, the dissertation aims to introduce a continuous-time representation of activity participation in tour-based model systems in practice. Activity-based travel demand forecasting model systems in practice today are largely tour-based model systems that simulate individual daily activity-travel patterns through the prediction of day-level and tour-level activity agendas. These tour level activity-based models adopt a discrete time representation of activities and sequence the activities within tours using rule-based heuristics. An alternate stream of activity-based model systems mostly confined to the research arena are activity scheduling systems that adopt an evolutionary continuous-time approach to model activity participation subject to time-space prism constraints. In this research, a tour characterization framework capable of simulating and sequencing activities in tours along the continuous time dimension is developed and implemented using readily available travel survey data. The proposed framework includes components for modeling the multitude of secondary activities (stops) undertaken as part of the tour, the time allocated to various activities in a tour, and the sequence in which the activities are pursued.

Second, the dissertation focuses on the implementation of a vehicle fleet composition model component that can be used not only to simulate the mix of vehicle types owned by households but also to identify the specific vehicle that will be used for a specific tour. Virtually all of the activity-based models in practice only model the choice of mode without due consideration of the type of vehicle used on a tour. In this research effort, a comprehensive vehicle fleet composition model system is developed and implemented. In addition, a primary driver allocation model and a tour-level vehicle type choice model are developed and estimated with a view to advancing the ability to track household vehicle usage through the course of a day within activity-based travel model systems. It is envisioned that these advances will enhance the fidelity of activity-based travel model systems in practice.