Taxi Trip Data Research Articles

Exploring spatio-temporal impact of COVID-19 on citywide taxi demand: A case study of New York City.

Coronavirus disease 2019 (COVID-19) has brought dramatic changes in our daily life, especially in human mobility since 2020. As the major component of the integrated transport system in most cities, taxi trips represent a large portion of residents' urban mobility. Thus, quantifying the impacts of COVID-19 on city-wide taxi demand can help to better understand the reshaped travel patterns, optimize public-transport operational strategies, and gather emergency experience under the pressure of this pandemic. To achieve the objectives, the Geographically and Temporally Weighted Regression (GTWR) model is used to analyze the impact mechanism of COVID-19 on taxi demand in this study. City-wide taxi trip data from August 1st, 2020 to July 31st, 2021 in New York City was collected as model's dependent variables, and COVID-19 case rate, population density, road density, station density, points of interest (POI) were selected as the independent variables. By comparing GTWR model with traditional ordinary least square (OLS) model, temporally weighted regression model (TWR) and geographically weighted regression (GWR) model, a significantly better goodness of fit on spatial-temporal taxi data was observed for GTWR. Furthermore, temporal analysis, spatial analysis and the epidemic marginal effect were developed on the GTWR model results. The conclusions of this research are shown as follows: (1) The virus and health care become the major restraining and stimulative factors of taxi demand in post epidemic era. (2) The restraining level of COVID-19 on taxi demand is higher in cold weather. (3) The restraining level of COVID-19 on taxi demand is severely influenced by the curfew policy. (4) Although this virus decreases taxi demand in most of time and places, it can still increase taxi demand in some specific time and places. (5) Along with COVID-19, sports facilities and tourism become obstacles on increasing taxi demand in most of places and time in post epidemic era. The findings can provide useful insights for policymakers and stakeholders to improve the taxi operational efficiency during the remainder of the COVID-19 pandemic.

Investigation of traffic-driven epidemic spreading by taxi trip data

Urban transportation systems account for tremendous population movements, which can also trigger epidemic outbreaks. This paper investigates the coupling of epidemic spreading and human mobility via taxi-trip data analysis and the Markov chain approach, and proposes targeted epidemic prevention measures. Significant variations in travel patterns are observed across different taxi zones within the city, and these disparities have a substantial impact on epidemic dynamics. First, the probability of taxi drivers and passengers traveling among taxi zones is obtained from empirical data. Abstracting human travel pattern as Markov process, a traffic-driven epidemic spreading model is established. Considering the impact of trip probability on disease spreading, the model can effectively reproduce the outbreak of COVID-19 in New York City, with correct features in different boroughs. Quantitative parameters are derived to indicate the influence of taxi zones and origin-destination trips on epidemic transmission. Applying prevention measures to a small number of important zones or key origin-destination trips in the early stage of spreading, the scale of epidemic outbreaks can be significantly reduced. This research offers insights for suppressing epidemic spread in densely populated metropolitan areas, with the potential to benefit policy efforts.

Are ride-hailing services safer than taxis? A multivariate spatial approach with accommodation of exposure uncertainty

Despite many research efforts on ride-hailing services and taxis, limited studies have compared the safety performance of the two modes. A major challenge is the need for reliable mode-specific exposure data to model their safety outcomes. Moreover, crash frequencies of the two modes by injury severities tend to be spatially and inherently correlated. To fully address these issues, this study proposes a novel multivariate conditional autoregressive model considering measurement errors in mode-specific exposures (MVCARME). More specially, a classical measurement error structure accommodates the uncertainty of estimated mode-specific exposures, and a multivariate spatial specification is adopted to capture potential spatial and inherent correlations. The model estimation is accelerated by an integrated nest Laplace approximation method. The census tracts in the city of Chicago are set as the spatial analysis unit. The mode-specific exposures (vehicle-mile-traveled) in each census tract are estimated by trip assignments using ride-hailing and taxi trip data in 2019. The modeling results indicate that both ride-hailing crashes and taxi crashes are positively associated with transportation factors (e.g., vehicle-mile-traveled, mode-specific vehicle-mile-traveled, and traffic signal numbers), land use factors (i.e., number of educational and alcohol-related sites), and demographic factors (e.g., median household income, transit ratio, and walk ratio). By comparison, the proposed model outperforms the others (i.e., negative binomial models and multivariate conditional autoregressive model) by yielding the lowest deviance information criterion (DIC), Watanabe-Akaike information criterion (WAIC), mean absolute error (MAE), and root-mean-square error (RMSE). According to the results of t-tests, ride-hailing services are found to be prone to a higher risk of minor injury crashes compared with taxis, despite no significant difference between the risks of severe injury crashes. Methodologically, this study adds a robust safety evaluation approach for comparing crash risks of different modes to the literature. At the same time, practically, it provides researchers, practitioners, and policy-makers insights into the safety management of various mobility alternatives.

Revealing intra-urban hierarchical spatial structure through representation learning by combining road network abstraction model and taxi trajectory data

ABSTRACT The unprecedented urbanization in China has dramatically changed the urban spatial structure of cities. With the proliferation of individual-level geospatial big data, previous studies have widely used the network abstraction model to reveal the underlying urban spatial structure. However, the construction of network abstraction models primarily focuses on the topology of the road network without considering individual travel flows along with the road networks. Individual travel flows reflect the urban dynamics, which can further help understand the underlying spatial structure. This study therefore aims to reveal the intra-urban hierarchical spatial structure by integrating the road network abstraction model and individual travel flows. To achieve this goal, we 1) quantify the spatial interaction relatedness of road segments based on the Word2Vec model using large volumes of taxi trip data, then 2) characterize the road abstraction network model according to the identified spatial interaction relatedness, and 3) implement a community detection algorithm to reveal sub-regions of a city. Our results reveal three levels of hierarchical spatial structures in the Wuhan metropolitan area. This study provides a data-driven approach to the investigation of urban spatial structure via identifying traffic interaction patterns on the road network, offering insights to urban planning practice and transportation management.

Identifying Important Nodes in Trip Networks and Investigating Their Determinants.

Describing travel patterns and identifying significant locations is a crucial area of research in transportation geography and social dynamics. Our study aims to contribute to this field by analyzing taxi trip data from Chengdu and New York City. Specifically, we investigate the probability density distribution of trip distance in each city, which enables us to construct long- and short-distance trip networks. To identify critical nodes within these networks, we employ the PageRank algorithm and categorize them using centrality and participation indices. Furthermore, we explore the factors that contribute to their influence and observe a clear hierarchical multi-centre structure in Chengdu's trip networks, while no such phenomenon is evident in New York City's. Our study provides insight into the impact of trip distance on important nodes within trip networks in both cities and serves as a reference for distinguishing between long and short taxi trips. Our findings also reveal substantial differences in network structures between the two cities, highlighting the nuanced relationship between network structure and socio-economic factors. Ultimately, our research sheds light on the underlying mechanisms shaping transportation networks in urban areas and offers valuable insights into urban planning and policy making.

New York City Taxi Trip Duration Prediction Using Machine Learning

Abstract: Given the complexity of urban transportation networks and the multiple variables that might affect journey times, estimating the length of cab rides in New York City (NYC) is a difficult process. In this study, we provide a unique method for resolving this issue that makes use of machine learning techniques and a wide range of attributes gleaned from taxi trip data. We start by gathering a sizable collection of historical records of NYC taxi trips,providing specifics like pick-up and drop-off points, timestamps, and lengths of trips. To deal with outliers, missing values, and geographical and temporal irregularities, we preprocess the data. Furthermore, we design a broad range of characteristics, such as geographic coordinates, time of day, day of the week, and weather conditions, to capture the spatial, temporal, and contextual elements of each journey. Then, using gradient boosting methods, we create a prediction model that efficiently uncovers the intricate patterns seen in the data. We carefully adjust the model's hyperparameter to enhance performance and use cross-validation techniques to guarantee resilience. In addition, we apply ensemble techniques to enhance prediction precision and minimise model bias. We conduct lengthy tests on a held-out test set and compare the performance of our model to a number of baseline techniques frequently employed in triptime prediction in order to assess the efficacy of our suggested strategy. The outcomes show that our strategy works better than the competition, with lower prediction errors and higher accuracy. We also do interpretability assessments to learn more about the variables that have the most impact on estimates of trip time. Our results demonstrate the potential of feature engineering and M L approaches for precise and trustworthy taxi trip length prediction in NYC. The suggested method not only helps taxi service companies by allowing them to more accurately predict journey lengths, but it also improves customer experience by giving more precise travel time estimates. Additionally, our approach may be used as a starting point for future studies in the field of urban transportation prediction, enabling better efficiency and planning in urban mobility networks.

QR분해를 사용한 RHadoop의 맵리듀스 기반 다중선형회귀모형 추정 알고리즘

RHadoop enables R users to perform big data analytics under R programming environment by integrating R with Hadoop that supports distributed storing and parallel processing of large-scaled data. This article proposes a RHadoop-based MapReduce programming model for estimating multiple linear regression models utilizing QR factorization. Our proposed algorithm accommodates the most common type of big data that has a vast number of data points with only a few hundred variables. For QR factorization over massive-scaled data, our algorithm employs DirectQR method proposed by Benson, Gleich, Demmel(2013); however, it does not necessitate its iterative steps to estimate regression coefficients. Through a comparative simulation study, our algorithm is compared with a MapReduce-based algorithm proposed in the previous studies. For generating realistic synthetic data, we utilize NYC(New York City) yellow taxi trip data reported to NYC Taxi and Limousine Commission. In the simulation, we measure estimation time and accuracy of each algorithm under several assumptions with respect to the strength of association between independent variables and the noise level of error terms in the regression models.

Geostatistics on Real-Time Geodata Streams—An Extended Spatiotemporal Moran’s I Index with Distributed Stream Processing Technologies

The availability of geodata with high spatial and temporal resolution is increasing steadily. Often, these data are continuously generated by distributed sensor networks and provided as geodata streams. Geostatistical analysis methods, such as spatiotemporal autocorrelation, have thus far been applied primarily to historized data. As such, the advantages of continuous and up-to-date acquisition of geodata have not yet been transferred to the analysis phase. At the same time, open-source frameworks for distributed stream processing have been developed into powerful real-time data processing tools. In this paper a methodology is developed to apply analyses of spatiotemporal autocorrelation directly to geodata streams through a distributed streaming process using open-source software frameworks. For this purpose, we adapt the extended Moran’s I index for continuous and up-to-date computation, then apply it to simulated geospatial data streams of recorded taxi trip data. Various application scenarios for the developed methodology are tested and compared on a distributed computing cluster. The results show that the developed methodology can provide geostatistical analysis results in real time. This research demonstrates how modern datastream processing technologies have the potential to significantly advance the way geostatistical analysis can be performed and used in the future.

Two-Sided Deep Reinforcement Learning for Dynamic Mobility-on-Demand Management with Mixed Autonomy

Autonomous vehicles (AVs) are expected to operate on mobility-on-demand (MoD) platforms because AV technology enables flexible self-relocation and system-optimal coordination. Unlike the existing studies, which focus on MoD with pure AV fleet or conventional vehicles (CVs) fleet, we aim to optimize the real-time fleet management of an MoD system with a mixed autonomy of CVs and AVs. We consider a realistic case that heterogeneous boundedly rational drivers may determine and learn their relocation strategies to improve their own compensation. In contrast, AVs are fully compliant with the platform’s operational decisions. To achieve a high level of service provided by a mixed fleet, we propose that the platform prioritizes human drivers in the matching decisions when on-demand requests arrive and dynamically determines the AV relocation tasks and the optimal commission fee to influence drivers’ behavior. However, it is challenging to make efficient real-time fleet management decisions when spatiotemporal uncertainty in demand and complex interactions among human drivers and operators are anticipated and considered in the operator’s decision making. To tackle the challenges, we develop a two-sided multiagent deep reinforcement learning (DRL) approach in which the operator acts as a supervisor agent on one side and makes centralized decisions on the mixed fleet, and each CV driver acts as an individual agent on the other side and learns to make decentralized decisions noncooperatively. We establish a two-sided multiagent advantage actor-critic algorithm to simultaneously train different agents on the two sides. For the first time, a scalable algorithm is developed here for mixed fleet management. Furthermore, we formulate a two-head policy network to enable the supervisor agent to efficiently make multitask decisions based on one policy network, which greatly reduces the computational time. The two-sided multiagent DRL approach is demonstrated using a case study in New York City using real taxi trip data. Results show that our algorithm can make high-quality decisions quickly and outperform benchmark policies. The efficiency of the two-head policy network is demonstrated by comparing it with the case using two separate policy networks. Our fleet management strategy makes both the platform and the drivers better off, especially in scenarios with high demand volume. History: This paper has been accepted for the Transportation Science Special Issue on Emerging Topics in Transportation Science and Logistics. Funding: This work was supported by the Singapore Ministry of Education Academic Research [Grant MOE2019-T2-2-165] and the Singapore Ministry of Education [Grant R-266-000-135-114].

Development of predictive modeling and deep learning classification of taxi trip tolls

Several studies discussed the predictive modeling of deep learning in different applications such as classifying tissue features from microstructural data, Crude Oil Prices, mechanical constitutive behavior of materials, microbiome data, and mineral prospectively. Commercial navigation includes a wealth of trip-related data, including distance, expected journey time, and tolls that may be encountered along the way. Using a classification algorithm, it is possible to extract drop-off and pickup locations from taxi trip data and estimate if the tour would incur tolls. In this work, let’s use the classification learner to create classification models, compare their performance, and export the findings for additional study. The workflow for the classification learner is the same as for the regression learner. The purpose is to make predictions based on fresh data in order to see how well the model performs with new data. To train the model, it’s critical to separate the data set. The combined training and validation data is next pre-processed, which involves tasks such as cleaning and developing new features skills. Once the data has been prepared, it’s time to begin the supervised machine learning process and test a number of ways to identify the best model, such as the type of model that should be used, the important features, and the best parameters of the model to find the best fit for the considered data. The results of analyzing different predictive multiclass classification models with taxi trip tolls show that it is possible to use a machine learning-based model when we like to avoid road tolls depending on historical data on taxi trip tolls. The outcome of this study can help to expect road tolls from the drop-off and pickup locations of a taxi data

Taxi Demand Forecasting Utilizing Neighbourhood Influence Based on Ensemble Learning

Abstract: This Time series forecasting (TSF) assists in making better strategic decisions under uncertain circumstances so that financial crisis can be avoided, wise investments can be made, under/over contracting of utility can be avoided, staffs can be scheduled appropriately, service providers can provide better service, mankind can get prepared for natural disasters and many more. However, the accuracy in forecasting plays a vital role and achieving such is a challenging task owing to the vagueness and nonlinearity associated with most of the real world time series. Therefore, improving the forecasting accuracy has become a keen area of interest among the forecasters from different domains of science and engineering. In this work, a time series forecasting model for Chicgo taxi trip data is proposed. The proposed model utilizes neighbourhood influence in terms of taxi demand to predict taxi demand. Keywords: Time series forecasting, Regularization, Regression

Exploring built environment factors that influence the market share of ridesourcing service

The ridesourcing service has taken a large portion of the taxi market in the past few years. Many studies have explored the influencing factors of traditional taxi and ridesourcing demand in different areas of the city. Few studies investigated the market share of ridesourcing in different regions of the city. Understanding the spatial distribution of the market share of ridesourcing service and the factors that determine this distribution could help government agencies understand the role that each type of service plays in the transportation system and evaluate the effect of different public policies on the two types of services. This paper studies this topic by constructing a panel fractional regression model based on the taxi trip data of New York City in 2017. Results show that the market share of the traditional taxi is declining while that of ridesourcing service is increasing. The market share of ridesourcing services is higher in remote areas, areas with low population density, low density of transportation facilities, low household income, and high proportion of young residents. It indicates that ridesourcing service could provide more equitable service by better serving the underserved areas while traditional taxi caters more to the high-demand areas.

A chance-constrained dial-a-ride problem with utility-maximising demand and multiple pricing structures

The classic Dial-A-Ride Problem (DARP) aims at designing the minimum-cost routing that accommodates a set of user requests under constraints at an operations planning level, where users’ preferences and revenue management are often overlooked. In this paper, we present a mechanism for accepting/rejecting user requests in a Demand Responsive Transportation (DRT) context based on the representative utilities of alternative transportation modes. We consider utility-maximising users and propose a mixed-integer programming formulation for a Chance Constrained DARP (CC-DARP), that captures users’ preferences via a Logit model. We further introduce class-based user groups and consider various pricing structures for DRT services. A customised local search based heuristic and a matheuristic are developed to solve the proposed CC-DARP. We report numerical results for both DARP benchmarking instances and a realistic case study based on New York City yellow taxi trip data. Computational experiments performed on 105 benchmarking instances with up to 96 nodes yield average profit gaps of 2.59% and 0.17% using the proposed local search heuristic and matheuristic, respectively. The results obtained on the realistic case study reveal that a zonal fare structure is the best strategy in terms of optimising revenue and ridership. The proposed CC-DARP formulation provides a new decision-support tool to inform on revenue and fleet management for DRT systems on a strategic planning level.

Ridesharing and alcohol-related assaults in NYC: A spatial ecological case-crossover study

ObjectiveRidesharing has changed urban transportation and the distribution of some health outcomes, including alcohol consumption. Studies relating ridesharing to crime and violence at low space-time resolution (e.g., county-months) find mixed results. The aim of this study was to examine whether ridesharing was associated with increased incidence of alcohol-related assaults within highly resolved space-time units. MethodsThis spatial ecological case-crossover study used rideshare and taxi trip data from the New York City (NYC) Taxi and Limousine Commission for 2017–2018 and assault data from the NYC Police Department, aggregated within taxi zone-hours. Conditional logistic regression models estimated the odds of observing an assault for case taxi zone-hours in which an assault occurred compared to two control units of the same taxi zone-hour one week before (−168 h) and one week after (+168 h) relative to the number of rideshare trips. Separate analyses assessed assaults occurring at bars and restaurants. ResultsFrom 2017–2018, there were 47,124 nighttime assaults in the 262 taxi zones. There were 2482 taxi zone-hours at a bar and 693 taxi zone-hours at a restaurant that contained at least one nighttime assault. Ridesharing was positively associated with nighttime assaults at bars (OR: 1.050; 95% CI: 1.002–1.100) but not at restaurants (OR: 1.049; 95% CI: 0.943–1.168). ConclusionsAdditional ridesharing trips are associated with increased incidence of assaults at on-premise alcohol outlets in NYC at the precise hour and taxi zone of trip origins.

Spatial Origin-Destination Flow Imputation Using Graph Convolutional Networks

Due to the limitation of data collection techniques and privacy issues, the problem of missing spatial origin-destination flows frequently occurs. Data imputation provides great support for the acquisition of complete flow data, which enables us to better understand regional connections and mobility patterns. However, existing models or approaches neglect the network structure of spatial flows, thus resulting in inappropriate estimates and a low performance. The development of graph neural networks offers a powerful tool to deal with graph-structured data. In this article, we proposed a spatial interaction graph convolutional network model, which combines graph convolution and a mapping function to predict flow data from the perspective of network learning. This model utilizes geographical unit embedding in local spatial networks to improve prediction accuracy. A negative sampling technique is adopted to reduce misestimation. Experiments on Beijing taxi trip data verified the usefulness of our model in spatial flow prediction. We also demonstrated that a biased training sample had a negative impact on the model’s performance. More attributes of geographical units, a more proper negative sampling rate and a larger training set can increase the prediction accuracy of flow data.

Evaluating rail transit's comparative advantages in travel cost and time over taxi with open data in two U.S. cities

This paper presents a comparative analysis of rail transit and taxi by travel cost and time based on the large-scale taxi trip data and public transit schedule information in two major U.S. cities. To quantify the relative advantage of one mode over the other, we introduce the notion of travel gradient, which is travel cost difference divided by travel time difference. Based on the signs of travel cost and time differences, we classify all trips into four quadrants. Quadrant II trips are selected for further analysis because rail transit is identified to be competitive with taxi for such trips. We also explore the relation between various trip characteristics and travel gradient with and without considering the spatial variation of such a relation. Main research findings include: (1) around 70% of the taxi trips in the considered datasets can be substituted with rail transit trips if the maximum walking distance is 0.5 miles at each trip end; (2) for around 10% of taxi trips with both modes being viable, rail transit dominates taxi in both travel cost and time; for the rest, rail transit is competitive with taxi; (3) the marginal travel cost saving due to mode switching from taxi to rail transit is about $70; and (4) there exist clearly spatial variations of the relation between trip characteristics and travel gradient. The main policy recommendation from this study is that rail transit can be better marketed by highlighting its relative advantage over taxi in travel time and cost, especially for travels in certain directions and time periods.

Revealing spatiotemporal travel demand and community structure characteristics with taxi trip data: A case study of New York City.

Urban traffic demand distribution is dynamic in both space and time. A thorough analysis of individuals’ travel patterns can effectively reflect the dynamics of a city. This study aims to develop an analytical framework to explore the spatiotemporal traffic demand and the characteristics of the community structure shaped by travel, which is analyzed empirically in New York City. It uses spatial statistics and graph-based approaches to quantify travel behaviors and generate previously unobtainable insights. Specifically, people primarily travel for commuting on weekdays and entertainment on weekends. On weekdays, people tend to arrive in the financial and commercial areas in the morning, and the functions of zones arrived in the evening are more diversified. While on weekends, people are more likely to arrive at parks and department stores during the daytime and theaters at night. These hotspots show positive spatial autocorrelation at a significance level of p = 0.001. In addition, the travel flow at different peak times form relatively stable community structures, we find interesting phenomena through the complex network theory: 1) Every community has a very small number of taxi zones (TZs) with a large number of passengers, and the weighted degree of TZs in the community follows power-law distribution; 2) As the importance of TZs increases, their interaction intensity within the community gradually increases, or increases and then decreases. In other words, the formation of a community is determined by the key TZs with numerous traffic demands, but these TZs may have limited connection with the community in which they are located. The proposed analytical framework and results provide practical insights for urban and transportation planning.

CARE-Share: A Cooperative and Adaptive Strategy for Distributed Taxi Ride Sharing

Given the fast growth of on-demand transportation services and ride-sharing platforms, the concept of private vehicle ownership is rapidly declining. Although there are multiple fully-grown ride-sharing systems, they are proprietary and centrally controlled. Facilitating ride-sharing using a localized distributed coordination between the riders and the drivers is in need. However, fully distributed systems deal with a large number of variables and objectives and are often sub-optimal. In this paper, we propose a distributed ride-sharing system with multiple objectives which are often conflicting to each other. Therefore, we model it as a multi-objective optimization problem and solve it using the Ant Colony optimization technique which sports a multi-agent behavior. We critically analyze the spatio-temporal challenges posed by the ride sharing problem and define novel performance metrics to capture the underlying subtlety of the distributed system performance. An in-depth experimentation with recent large-scale single-ride taxi trip data from Chicago shows that our solution can ensure up to 79.65% success rate of ride sharing. We have shown that ride sharing is more successful during non-peak traffic hours due to less contention and a healthy balance in passenger and taxi numbers. Further, it has been observed that ride-sharing always reduces the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">total distance travelled</i> by all the taxis and the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">total number of taxis</i> on-road; both of which positively impact road congestion and environment. The results obtained from the experiments are very much comparable to real time behaviour of taxi networks. Finally, a revenue framework is proposed to analyse nuances of the operating environment.

Characterizing the urban spatial structure using taxi trip big data and implications for urban planning

Urban spatial structure is an important feature for assessing the effects of urban planning. Quantifying an urban spatial structure cannot only help in identifying the problems with current planning but also provide a basic reference for future adjustments. Evaluation of spatial structure is a difficult task for planners and researchers and this has been usually carried out by comparing different land use structures. However, these methods cannot efficiently reflect the influence of human activities. With the wide application of big data, analyzing data on human travel behavior has increasingly been carried out to reveal the relationship between urban spatial structure and urban planning. In this study, we constructed a human-activity space network using the taxi trip big data. Clustering at different scales revealed the hierarchy and redundancy of the spatial structure for assessing the appropriateness and shortcomings of urban planning. This method was applied to a case study based on one-month taxi trip data of Dongguan City. Existing urban spatial structures at different scales were retrieved and utilized to assess the effectiveness of the master plan designed for 2000 to 2015 and 2008 to 2020, which can help identify the limitations and improvements in the spatial structure designed in these two versions of the master plan. We also evaluated the potential effect of the master plan designed for 2016 to 2035 by providing a reference for reconstructing and optimizing future urban spatial structure. The analysis demonstrated that the taxi trip data are important big data on social spatial perception, and taxi data should be used for evaluating spatial structures in future urban planning.

Data-driven Distributionally Robust Optimization For Vehicle Balancing of Mobility-on-Demand Systems

With the transformation to smarter cities and the development of technologies, a large amount of data is collected from sensors in real time. Services provided by ride-sharing systems such as taxis, mobility-on-demand autonomous vehicles, and bike sharing systems are popular. This paradigm provides opportunities for improving transportation systems’ performance by allocating ride-sharing vehicles toward predicted demand proactively. However, how to deal with uncertainties in the predicted demand probability distribution for improving the average system performance is still a challenging and unsolved task. Considering this problem, in this work, we develop a data-driven distributionally robust vehicle balancing method to minimize the worst-case expected cost. We design efficient algorithms for constructing uncertainty sets of demand probability distributions for different prediction methods and leverage a quad-tree dynamic region partition method for better capturing the dynamic spatial-temporal properties of the uncertain demand. We then derive an equivalent computationally tractable form for numerically solving the distributionally robust problem. We evaluate the performance of the data-driven vehicle balancing algorithm under different demand prediction and region partition methods based on four years of taxi trip data for New York City (NYC). We show that the average total idle driving distance is reduced by 30% with the distributionally robust vehicle balancing method using quad-tree dynamic region partitions, compared with vehicle balancing methods based on static region partitions without considering demand uncertainties. This is about a 60-million-mile or a 8-million-dollar cost reduction annually in NYC.