Impacts of COVID-19 Travel Restrictions on Paratransit Users in Honolulu, Hawaii
Karl Kim, University of Hawaii Genesis Leong, University of Hawaii Eric Yamashita, University of Hawaii Jiwnath Ghimire, University of Hawaii
Show Abstract
Using comprehensive data on 48,000 paratransit trips in the City and County of Honolulu, following the initial onset of COVID-19, changes in travel behavior of paratransit users are described. Honolulu has one of the largest paratransit systems in the nation. The analysis provides insight into the effects of Covid-19 on paratransit users. In addition to a sharp reduction in total paratransit trips (63%), non-essential trips fell by a greater proportion (96 percent). Overall medical trips decreased by 69.23%, yet trips for dialysis increased by 3.76%,while other health-related service trips fell by 66.3%. Other trip purposes including church, education, daycare, work, and personal are examined for selected weeks to compare representative day by day and week by week changes. The declines in trip-making demonstrate not just a reduction in transportation services to vulnerable populations but also diminished support of the functioning, wellbeing, and quality of life among paratransit users.Trip destinations were mapped and analysed as to clusters which could be compared to disease outbreak hotspots or locations with large public gatherings where the risks of contagion could be higher. While the full impacts of the travel restrictions and the pandemic have yet to be accounted for, efforts to maintain and improve paratransit services require flexible, adaptive, and innovative approaches during and after the health crisis. Strategies for the protection of users and operators and continued services include improved sanitation, continued testing and monitoring of operators and users, better report and data sharing systems, and continued commitment to innovation and service quality.
|
TRBAM-21-04331
|
Intercity Passenger Transportation Plan in a Brazilian State
LUIZ GOMES, GISTRAN
Show Abstract
An intercity passenger transportation master plan has been developed for the Brazilian State of Mato Grosso do Sul, whose main findings and proposals are reported in the present paper. This is a typical rural State, its economy relying substantially on agriculture (grains and meat products). Its demographic density is only 7,78 inhabitants per square kilometer, being the sixth major state in territorial area.
The main objectives of the master plan aims to rationalize and modernize the operation of the services provided in the next to 180 routes, running under the regime of public concessions. As a result of comprehensive surveys, demand analysis and modeling, a proposal was carried out for an aggregation of routes around the nine main regions of the State (including Pantanal region).
The plan´s main innovation is the proposal of some flexible routes that will work on demand transportation to be operated by minibus. These vehicles have been operating short routes, authorized by the State Regulatory Agency, with a positive user experience evaluation, as found out in specific surveys made during the plan`s initial phase of development.
As a complement, a brief summary of new trends concerning the future of interstate routes is presented, in order to give a better understanding of the evolving situation of bus intercity services in Brazil. A new policy was recently announced by Federal Government, which in a recent decree allows the access for new operators, provided that they are able to attend new transportation demands.
|
TRBAM-21-00572
|
Aberrant Driving Behavior Analysis and Driving Risk Level Assessment for Intercity Bus Fleets
Chien-Hung Wei, National Cheng Kung University Yu-Chin Chen ( t90175@cc.kyu.edu.tw), Kao Yuan University Jia-Wei Lai, National Cheng Kung University
Show Abstract
With the growing use of public transportation, intercity busses play an essential role around the world. However, traffic accidents involving busses commonly result in serious casualties and financial loss. In recent years, the fatality rate of busses has been ten times that of sedans, demonstrating the urgent need to improve bus safety in Taiwan. Thus, intercity bus carriers are increasingly valuing driving behavior. Previous studies showed that bus drivers’ behavior is significantly influenced by their personal human factors and their driving risk based on aberrant driving behavior can be used as an indicators of traffic accidents. The aim of this research was to was to develop a driving risk level evaluation method to measure aberrant driving behavior, and to quantify relative driving risk by applying artificial neural network models to predict the risk level of each driver based on individual human factors. The empirical results showed that the predictive models had a high accuracy. The Spearman correlation coefficient was then used to analyze the correlation between the human factors and driving risks. Driving fatigue, symptoms, disease, and high neuroticism were found to cause high driving risks, while adequate sleep hours, high agreeableness, and high annual household income led to low driving risks. By establishing a systematic driving risk assessment mechanism, the intercity bus industry can reduce the frequency of traffic accidents and contribute to public transportation management.
|
TRBAM-21-01172
|
Optimization of Intercity Demand Response Transit Service Design Considering Zoning Strategy
Hongtai Yang, Southwest Jiaotong University Rong Zheng ( xzwtzr@163.com), Southwest Jiaotong University Wenbo Fan, Southwest Jiaotong University Xiaojian Zhang, University of Florida Zhaolin Zhang, Southwestern University of Finance and Economics
Show Abstract
Intercity Demand Response Transit (IDRT) is a type of “door-to-door” intercity transit that picks up passengers from one city and transport them to another city in a demand responsive way. Despite its rapid development in recent years, its service design has not been studied yet. This study proposes a method to design the optimal IDRT service, which is to find the optimal service area of the satellite city, number of zones of the main city, service area of each zone, and headway. The objective is to minimize the generalized system cost that consists of transit agency's operation cost, passengers' waiting and in-vehicle traveling cost. A two-stage algorithm is developed by combining the greedy algorithm, multiplier method and sensitivity coordination optimization method. The effectiveness of the proposed model and solution algorithm is verified using simulation. Sensitivity analyses are carried out with respect to demand densities, service areas, line-haul distance, and vehicle capacities. The findings in this study could be used to guide the design and operation of IDRT in practice.
|
TRBAM-21-04239
|
INTERCITY COMMUTER BUSES IN THE U.S.A. – SHARED MOBILITY CONGESTION MITIGATION SOLUTIONS CAUGHT IN A GLOBAL PANDEMIC AND A COMPLEX REGULATORY WEB
Matthew Daus, City College of New York
Show Abstract
This article provides an overview of the trends and future pathways for intercity commuter buses and explains various laws and regulations governing these services in major United States cities. The article also provides considerations for companies when expanding service or adding a stop in a new market following the COVID-19 pandemic, which has had major ramifications on how people travel.
Keywords: Intercity Bus, Commuting, COVID-19
|
TRBAM-21-03227
|
Measuring Benefits of Rural and Small Urban Transit in Greater Minnesota
Jeremy Mattson, North Dakota State University Del Peterson, North Dakota State University
Show Abstract
The objective of this research was to measure the benefits of rural and small urban transit services in Minnesota. The study accomplished this by first identifying, describing, and classifying the potential benefits of transit. Second, a method was developed for measuring these benefits. Where possible, benefits were quantified in dollar values. Other benefits that could not be monetized were either quantified in another way or described qualitatively. The study included an analysis of societal benefits and economic impacts within local communities. Third, the developed method was applied to a series of six case studies across Greater Minnesota. Data were collected through onboard rider surveys for each of the six transit agencies. Total benefits and benefit-cost ratios were estimated for the six transit agencies—all showed benefits that exceeded costs—and results were generalized to Greater Minnesota. Economic impacts were also estimated showing impacts on jobs, labor income, and value added. This research provides information to assess the benefits of public spending on transit, which gives decision makers the data needed to inform investment decisions.
|
TRBAM-21-00197
|
Time-varying Accessibility to Senior Community Centers by Public Transit in Philadelphia
Shengxiao Li ( lsx@design.upenn.edu), University of Pennsylvania Anna Duan, University of Pennsylvania Haoyu Hu, Peking University
Show Abstract
We examined the time-varying accessibility to senior community centers via public transit in each census tract of Philadelphia, a city in the United States and how accessibility varies across census tracts with different shares of seniors and seniors in the marginalized groups using GTFS and socio-economic data from the Census Bureau of the States. We found that accessibility to senior centers declines from the center city to the suburbs, but accessibility varies temporally on weekdays and weekends, with the lowest accessibility occurring at 2 pm on weekdays. Some downtown census tracts on weekdays and some suburban census tracts on weekends have better accessibility at 6 pm when most senior community centers are closed. In terms of the relationship between accessibility and shares of seniors and vulnerable seniors at the census tract level, we found that accessibility to senior community centers is in positive proportion to the shares of seniors under the poverty line and with no cars in their household, while the accessibility is in a negative relationship with the share of seniors in each census tract. The study suggests that the transit agency and the Area Agency on Aging should work together to better match the operation hours of the transit system and community centers with the social needs of heterogeneous social groups among seniors. However, public transit is not the panacea to solve the transportation problems of the transport disadvantaged given the dispersed distribution of seniors of heterogeneous disadvantaged conditions.
|
TRBAM-21-00379
|
Continuous Approximation Model for Hybrid Flexible Transit Systems in Rural Areas
Charalampos Sipetas ( csipetas@umass.edu), University of Massachusetts, Amherst Eric Gonzales, University of Massachusetts, Amherst
Show Abstract
Flexible transit systems are a way to address challenges associated with conventional fixed route and fully demand responsive systems. Existing studies indicate that such systems are often planned and designed without established guidelines, and optimization techniques are rarely implemented on actual flexible systems. This study presents a hybrid transit system where the degree of flexibility can vary from a fixed route service (with no flexibility) to a fully flexible demand responsive transit system. Such a system is expected to be beneficial in areas where the best transit solution lies between the fixed route and fully flexible systems. Continuous approximation techniques are implemented to model and optimize the stop spacing on a fixed route corridor, as well as the boundaries of the flexible region in a corridor. Both user and agency costs are considered in the optimization process. A numerical analysis compares various service areas and demand densities using input variables with magnitudes similar to those of real-world case studies. Sensitivity analysis is performed for service headway, percent of demand served curb-to-curb, and user and agency cost weights in the optimization process. The hybrid system proposed here achieves estimated user benefits of up to 35% when compared to fixed route systems, under different case scenarios. Flexible systems are particularly beneficial for serving corridors with low or uncertain demand. This provides value for rural corridors as well as communities in which transit ridership has dropped significantly due to the COVID-19 pandemic.
|
TRBAM-21-02962
|
The State of Demand-Responsive Transit in Canada
Willem Klumpenhouwer, University of Toronto Amer Shalaby, University of Toronto Lee Weissling, Ontario Society of Professional Engineers
Show Abstract
Demand-responsive transit is a growing phenomenon in Canada, and is being increasingly considered as an alternative to fixed-route transit in lower density suburban and rural areas. This desire to provide coverage service at a lower cost is further driven by the introduction of application-based systems and cloud computing, which can provide more dynamically optimized routing and scheduling of services than was previously available. This research investigated the current state of demand-responsive transit service in Canada by studying transit agencies that have been involved in demand-responsive transit projects in the recent past, present, and near future. Through a series of interviews, a synthesis of the current state of practice in demand-responsive transit in Canada is provided. This includes discussions of planning considerations, operations, procurement, and evaluation. In general, transit agencies in Canada are thoughtfully optimistic about the potential of modern demand-responsive transit service, recognizing its potential without considering it a blanket solution to mobility in their regions.
|
TRBAM-21-00594
|
Research on Demand Analysis and Operation Organization of Flex-route Transit Based on Retrograde Driving Rule
Jin Zhang, Southeast University Wenquan Li, Southeast University Guoqing Wang, China Design Group Co., Ltd Jingcai Yu, Southeast University
Show Abstract
This paper focuses on the demand analysis and route planning of flex-route transit. Through mining multi-source traffic data, the potential users, travel demand, and selection of flex-route transit stations are analyzed, which provides the basis for the subsequent route planning. A flex-route transit route design strategy is proposed, the backtracking point selection rule is designed, and the route planning under the retrograde driving rule is carried out. A route planning model of flex-route transit is established, which takes the weighted sum of vehicle running time, waiting time and on-board time as the objective. And on this basis, the sensitivity analysis of the relevant factors in the retrograde driving rule is carried out. Through the case study, the validity of the route planning model under the retrograde driving rule is verified. The results demonstrate that the retrograde driving rule can effectively reduce the total driving distance of the vehicle and the various time cost of the system.
|
TRBAM-21-01820
|
Evaluation of the Federal Transit Administration’s Transit and Health Access Initiative: Case Highlights, Findings, and Recommendations
Jennifer Flynn, USF Center for Urban Transportation Research Victoria Perk, University of South Florida Austin Sipiora, University of South Florida
Show Abstract
To address transportation barriers to healthcare access, FTA provided funding for 19 Transit and Health Access demonstration projects. The goal of the grants was to find and test promising, replicable public transportation healthcare access solutions that support the “triple aim” of increased access to care, improved health outcomes, and reduced healthcare costs. Eight of the 19 projects were funded through 49 U.S.C. § 5312. All demonstration projects that receive assistance under Section 5312 are required by federal transportation law to undergo an independent evaluation. FTA engaged the Center for Urban Transportation Research (CUTR) to serve as the independent evaluator (IE) of these eight projects and to produce a final evaluation report. This paper highlights two compelling projects from the evaluation of the Transit and Health Access Initiative, selected for their innovative approach and wealth of lessons learned. After in-depth case descriptions of these two projects, the authors present the overall findings and recommendations that emerged from the full evaluation effort.Keywords: Public Transportation, Transit, Health, Healthcare, Barriers, Access, Mobility, Coordination, Human Services, Transit and Health Access Initiative, Demonstration, Pilot
|
TRBAM-21-01751
|
Criteria to Transfer Paratransit Trips into a Transit Network
Camille Garnier, Ecole Polytechnique de Montreal Martin Trépanier, Ecole Polytechnique de Montreal Catherine Morency, Ecole Polytechnique de Montreal
Show Abstract
Paratransit is an on-demand shared ride mobility service for disabled people. In the city of Montreal (Canada), the demand for this service is rapidly growing but unfortunately, resources are not. Therefore, the aim of the research is to develop and test criteria to identify paratransit trips that could be potentially transferred to the transit network. Four criteria are proposed to identify these trips. The first one is related to the universal accessibility level of each metro station and bus stop. The second one is based on the spatial location of both origin and destination trip ends and their proximity to the transit network. The third one is related to the autonomy level of paratransit users as they must be able to use the transit network by themselves. The last one relates to the time of travel as trips to transfer must not happen during the typical peak hours of the transit network. With these four criteria, it is possible to estimate how many paratransit trips could be considered as potentially transferable. In the current transit network of the STM (Société de transport de Montréal), using these criteria, some 14% of the paratransit trips would be considered transferable under 100-meter distance to transit stops threshold. Over one year, it means that some 500 000 trips could be done using the transit network instead of the paratransit service. This alternative can provide more flexible service to travelers as well as a cheaper way to provide travel options to the disabled community people.
|
TRBAM-21-00654
|
Shared Mobility in Los Angeles County: A Case Study in Westside Cities
Huajun Chai, University of California, Davis Caroline Rodier, University of California, Davis Ihab Kaddoura, Technische Universitat Berlin
Show Abstract
Shared autonomous vehicle (SAV) as a shared mobility has attracted a great amount of attention recently. SAVs have the potential to reduce private vehicle ownership, improve the accessibility to high-quality transportation for residents from rural and under-developed areas, and reduce vehicle miles traveled (VMT) and emissions. Especially, when it is coupled with public transit system as a first/last mile transport service, SAVs could encourage travelers to choose public transit over other travel modes. It is an effective approach to promote greener travel. However, due to lack of service, long waiting time from insufficient fleet size, inconvenience from ridesharing and many other factors, SAVs have many challenges to overcome. In this paper, a regionally calibrated large scale agent-based multi-modal traffic simulation model for Los Angeles County is developed and used to study the effects of demand response transport (DRT, a type of shared mobility services using SAVs) under two different policy, including congestion levels, private vehicle ridership, and public transit usages. We find that DRT services could reduce private vehicle ridership. Lower income group might foresee a greater increase in public transit usage compared with higher income groups due to the convenient and affordable access by DRT to public transit that offers cheap transportation compared with other alternative transportation modes like private vehicles, taxis, and pure DRT rides. Lower transit cost could further reduce private vehicle ridership and attract more public transit demand, including direct transit usage and transit coupled with DRT as a first/last mile transport mode.
|
TRBAM-21-01789
|
Simulation of Potential Use Cases for Shared Mobility Services in the City of Ann Arbor
Richard Twumasi-Boakye, Ford Motor Company Xiaolin Cai, Ford Motor Company James Fishelson, Ford Motor Company Andrea Broaddus, Ford
Show Abstract
In this paper, we model and simulate special use cases of on-demand shared mobility services for the City of Ann Arbor, Michigan. We define shared mobility as any motor-vehicle-served transportation option between private vehicles and public transit, such as taxis, demand-responsive transit, and dynamic shuttles. Here, we present and evaluate a suite of four different service types that could potentially complement existing transportation services in Ann Arbor. A novel aspect of this study is that it tests scenarios that were developed in consultation with city planners looking for insights on real-world problems. This study used fleet simulation software to test four service configuration scenarios for a hypothetical on-demand shared mobility service: city-wide shuttle, a corridor-based downtown shuttle, a Park & Ride shuttle, and transit-complementary service. Three levels of demand were tested for each scenario, 3, 9, and 15 percent of all private vehicle trips in the city. Findings indicate that city-wide on-demand shared mobility services struggle to achieve higher vehicle occupancies than private vehicles at approximately 1.4. Service configurations with aggregated trip density can result in slightly improved occupancies, as found in downtown and Park & Ride shuttle scenarios. More impactful is aggregating demand by moving from “many-to-many” routing as with city-wide floating services to “many-to-one” routing as with downtown or Park & Ride shuttle services, which can increase vehicle occupancy from 1.4 to almost 2. Lastly, we also discuss potential benefits in terms of reduced congestion and parking needs.
|
TRBAM-21-01942
|
Planning for the Integration of Ride-Hailing Services with Transit: The Case of Small Transit Agencies
V. Dimitra Pyrialakou ( dimitra.pyrialakou@mail.wvu.edu), West Virginia University Parisa Hajibabaee, West Virginia University Leily Farrokhvar, West Virginia University Amrit Williams, West Virginia University
Show Abstract
Ride-hailing programs established through partnerships between transit agencies and transportation network companies (TNCs) is an emerging innovative practice in public transportation services. Literature and practice attest to the potential of these programs. Recent research reports have produced how-to guides and other useful material that can guide transit agencies through the various stages involved in exploring, understanding, defining, and establishing such a program. However, these materials lack adequate representation of programs established within small transit agencies, because very few of the existing case studies are of ride-hailing programs operating in rural and small urban areas or small University/College/campus towns. In this study, to address this gap, we focus on the early planning stages of establishing a ride-hailing program within a small transit agency. We propose a framework based on multi-criteria decision analysis methods that can be used to explore the perceived and actual potential benefits, costs, and performance of different ride-hailing service models. We illustrate this methodology using the case study of Mountain Line Transit Authority, operating in Monongalia County, West Virginia. The results highlight the need for a local exploration of both perceived and actual measures of the benefits and costs of different service models. Concerning the empirical results, the findings universally suggest that any ride-hailing service model is perceived as an enhancement of the existing transit services. Finally, there is a consensus in that the area and the agency will be the least benefited from implementing a late night/early morning ride-hailing service model.
|
TRBAM-21-02917
|
Transportation Network Systems (TNCs) and Automated Vehicles (AVs): Bridging the Gap from City to Rural Areas
Selena Scott, Texas Southern University Gwendolyn Goodwin, Texas Southern University Teidra Darrett, Texas Southern University
Show Abstract
Transportation is critical to the economic vitality of large Metropolitan regions, smaller communities, and rural areas. Responding to transportation needs has been a priority in many states, including Texas. Emphasis has been placed on addressing increasing levels of traffic congestion, declining mobility, rural transit, and air quality concerns. The use of AVs and TNCs are being explored as ways to address these concerns. Multimodal approaches and the application of advanced technologies can be used to improve the movement of people and goods. The future is ultimately unknowable, but planning requires predictions of impending conditions, needs and possible solutions.
The purpose of this paper is to establish a framework for the need and use of AVs and TNCs within the rural public transit services. Subsequently this paper details the growth and abilities of autonomous vehicles and transportation network systems as well as establishes the need of effective and reliable public transit for areas outside city limits. In addition, the benefits, disadvantages and challenges of these ideals are discussed and reviewed. The resolution of the research is that AVs and TNCs for rural public transit are worth exploring and considering. TNCs and AVs are the future of transportation as they allow further reach for immobilized citizens and provide a safer, more sustainable traveling option that will directly contribute to citizens having better access to social and economic gains. In addition, AVs have the capability to bridge the gap between rural areas and the cities of which they correspond.
|
TRBAM-21-02006
|
Demand Responsive Transit Simulation of Wayne County, Michigan
Grace Kagho, Eidgenossische Technische Hochschule Zurich David Hensle, RSG Inc Miloš Balać, Eidgenossische Technische Hochschule Zurich Joel Freedman, RSG Richard Twumasi-Boakye, Ford Motor Company Andrea Broaddus, Ford James Fishelson, Ford Motor Company Kay Axhausen, Eidgenossische Technische Hochschule Zurich
Show Abstract
Abstract
Demand Responsive Transit (DRT) can provide an alternative to private cars and complement existing public transport services. DRT’s potential is enhanced by the advent of automation; such services are often referred to as shared autonomous vehicles (SAVs). However, the successful implementation of DRT services remains a challenge; as both researchers and policy makers can struggle to determine what sorts of places or cities are suitable for it. Research into car-dependent cities with poor transit accessibility are sparse. In this study, we address this problem, investigating the potential of DRT service in Wayne County, USA, whose dominant travel mode is a private car. Using an agent-based approach, we simulate DRT as a new mobility option for this region, thereby providing insights on its impact on operational, user, and system-level performance indicators. We test DRT scenarios for different fleet sizes, vehicle occupancy, and cost policies. The results show that a DRT service in Wayne County has certain potentials, especially to increase the mobility of lower-income individuals. However, we also show that introducing the service may slightly increase the overall VKT. Specific changes in service characteristics, like service area, pricing structure, or preemptive relocation of vehicles, might be needed to fully realize the potential of pooling riders in the proposed DRT service. We hope that this study serves as a starting point for understanding the impacts and potential benefits of DRT in Wayne County and similar low-density and car-dependent urban areas, as well as the service parameters needed for its successful implementation.
|
TRBAM-21-03155
|
An Overview of Data Collection Best Practices: Personal Mobility Gaps, Root Causes, and Solutions, for Older Americans Living in Rural Areas
Allun Ward, Prairie View A&M University Brandon Blond, Prairie View A&M University Judy Perkins, Prairie View A&M University
Show Abstract
The principal quest for conducting this research is to better understand personal mobility of older Americans living in rural areas. This is achieved through reviewing the methodologies utilized to collect data on mobility gaps and their root causes as well as discuss the trends that promote pursuing future personal mobility research. The information in this paper is a continuation to the first follow-up to research made possible by an on-going Toyota Mobility Challenge research project at Prairie View A&M University (PVAMU) which is located in Waller County, Texas.
Keywords: Personal Mobility, Older Americans, Rural Transportation
|
TRBAM-21-03399
|
Stochastic Optimization of Semi-Flexible Transit Operations
Show Abstract
Semi-flexible transit (SFT) is commonly discussed as a cost-effective alternative to serving public transportation users in low demand conditions. Despite its considerable potential, implementation of SFT is limited due two primary operating challenges: (a) fluctuating travel demand and (b) service unreliability. Most researchers recently are rigorously involved in developing complex algorithms and heuristics to handle operational planning issues while a very few focused on optimization of variables for SFT operation involving tactical decision making. Moreover, the optimization of decision variables are largely based on a single dimension of stochasticity, demand only. The present study proposes a methodology to optimize two decision variables, service headway and proportion of requests accepted for curb-to-curb service per trip while operating SFT following a route-deviation operating policy. Implementing stochasticity in both demand and vehicle arrival, we perform multi-objective optimization with two conflicting objectives as minimization of operator cost and user cost. Pertaining to vehicle delays and demand variability linked to values of decision variables in the Pareto set, we define the risks associated with selecting each value for attaining Pareto optimality. The risk is proportionate to occurrence of a decision variable value in the Pareto set. The study methodology can be adopted as a decision support tool to establish planning policies to optimize SFT operation while considering interests of both operator and user.
Keywords: Semi-flexible transit; Service headway; Curb-to-curb request; Operator cost; User cost; Multi-objective Optimization; Stochasticity
|
TRBAM-21-03454
|
Dynamic Ride-matching for Large-scale Transportation Systems
Taner Cokyasar ( tcokyasar@anl.gov), Argonne National Laboratory Felipe de Souza, Argonne National Laboratory Joshua Auld, Argonne National Laboratory Omer Verbas, Argonne National Laboratory
Show Abstract
Efficient dynamic ride-matching (DRM) in large-scale transportation systems is a key driver in transport simulations to yield answers to challenging problems. Although the DRM problem is simple to solve, it quickly becomes a computationally challenging problem in large-scale transportation system simulations. Therefore, in this study, we thoroughly examine the DRM problem dynamics and propose an optimization-based solution framework to solve the problem efficiently. To benefit from parallel computing and reduce computational times, the problem's network is divided into clusters utilizing a commonly-used unsupervised machine learning algorithm along with a linear programming model. Then, these sub-problems are solved using another linear program to finalize the ride-matching. At the clustering level, our framework allows users adjusting cluster sizes to balance the trade-off between the computational time savings and the solution quality deviation. A case study on the Chicago Metropolitan Area illustrates that our framework can reduce the average computational time by 58% at the cost of increasing the average pick up time by 26% compared to a system-optimum, i.e., non-clustered, approach. Another case study on a relatively small city, Bloomington, Illinois, shows that our framework provides quite similar results to the system-optimum approach in approximately 62% less computational time.
|
TRBAM-21-03909
|