Average Travel Time Research Articles

Adaptive Traffic Signal Control Method Based on Offline Reinforcement Learning

The acceleration of urbanization has led to increasingly severe traffic congestion, creating an urgent need for effective traffic signal control strategies to improve road efficiency. This paper proposes an adaptive traffic signal control method based on offline reinforcement learning (Offline RL) to address the limitations of traditional fixed-time signal control methods. By monitoring key parameters such as real-time traffic flow and queue length, the proposed method dynamically adjusts signal phases and durations in response to rapidly changing traffic conditions. At the core of this research is the design of a model named SD3-Light, which leverages advanced offline reinforcement learning to predict the optimal signal phase sequences and their durations based on real-time intersection state features. Additionally, this paper constructs a comprehensive offline dataset, which enables the model to be trained without relying on real-time traffic data, thereby reducing costs and improving the model’s generalization ability. Experiments conducted on real-world traffic datasets demonstrate the effectiveness of the proposed method in reducing the average travel time. Comparisons with several existing methods highlight the clear advantages of our approach in enhancing traffic management efficiency.

A multi-period capacitated facility location problem with maximum travel time and backup service for locating and sizing EMS stations

Emergency medical services (EMS) is a system that provides emergency medical care for incidents involving serious illness or injury. The location of EMS stations plays an essential role in delivering effective and efficient medical services. Numerous location models have been developed for locating and sizing EMS stations. However, it remains challenging to satisfy all EMS planning criteria within a single location model. In this study, a multi-period capacitated facility location problem with maximum travel time and backup service (EMSLSP) is proposed for locating and sizing EMS stations. The most important criteria for EMS planning are taken into account in EMSLSP: the demand changes due to population mobility, the maximum service capacity of an ambulance, the maximum number of ambulances at each EMS station, the maximum travel time from each EMS station to the locations it serves, the full coverage of dynamic demand, the minimum percent of population covered by EMS service in a specific travel time, and a backup station for each demand location in case of need. A case study in Zhengzhou, a large city in China, demonstrates that effective and efficient locations and sizes of EMS stations can be determined by solving the EMSLSP with various planning parameters. Compared with the existing EMS systems, the average ambulance travel time and the percentage of the population served are significantly improved. Simulations of ambulance scheduling confirm that the relocated and resized EMS stations perform better than those in the existing system. The evaluation-optimization-simulation method outlined in this paper provides a comprehensive and effective approach for EMS station planning.

Wardrop Optimal Networks

Optimal flow allocation in directed networks is often analyzed through two types of equilibria. The first, called user equilibrium, occurs when the travel times on all utilized routes are the same and shorter than every unused route. The second, called system optimum, minimizes the average travel time across all routes. These two concepts represent the optimality for individual users and for the network as a whole, respectively, and generally do not coincide. This paper introduces and examines networks where user equilibrium and system optimum are identical, termed Wardrop Optimal Networks. Moreover, we illustrate that these flows remain consistent under several important transformations as well as under uniform changes in their latency functions.

Spatial Configuration and Accessibility Assessment of Recreational Resources in Hainan Tropical Rainforest National Park

Recreational resources, fundamental to ecological experiences, are critical in balancing conservation with development. Effective ecotourism planning is especially vital for newly established protected areas such as the Hainan Tropical Rainforest National Park in China’s developing system of natural conservation areas. Targeting Hainan Tropical Rainforest National Park, this study applies nearest neighbor index, kernel density analysis, and exploratory spatial data analysis (ESDA) to study the spatial pattern of 274 recreational resource points. Results indicate a clustered spatial pattern with significant differences in resource density among municipalities. Specifically, 98% of these resources can be reached in 3 h, with an average travel time of 91 min, and cultural resources exhibit greater accessibility than natural resources. Natural resource availability and ethnic culture are major factors of resource distribution and accessibility. This research offers a theoretical basis and practical guidance for optimizing recreational resource allocation and promoting ecotourism in the park, contributing to the ongoing discussion of sustainable tourism development.

Performance Evaluation for the Expansion of Multi-Level Rail Transit Network in Xi’an Metropolitan Area: Empirical Analysis on Accessibility and Resilience

As the main form of new urbanization, the coordinated development of cities in metropolitan areas requires reliable and efficient rail transit skeleton support. However, in the rapid development of metropolitan areas, the layout and analysis of multi-level rail transit systems have a certain lag. Taking the Xi’an metropolitan area as an example, this study analyzes the comprehensive accessibility and resilience of the multi-level rail transit network, and proposes an expansion plan accordingly. The traffic analysis zone (TAZ) is divided by towns and streets, and the relationship between points of interest (POIs) and the regional average level is analyzed using DEA. The improved weighted average travel time model is built with the analysis results as regional weights; a site selection model based on multiple construction influencing factors is proposed, and four expansion plans, namely, economic optimal, environmental optimal, transport optimal, and integrated optimal, are designed. The peak passenger flow scenario and the “failure–reparation” scenario during the entire operation period are designed to analyze the resilience of four plans, and the resilience is quantified by the elasticity curve of the maximum connected subgraph ratio (MCSR) changing over time. The research results show that the transport optimal plan has the best comprehensive accessibility and resilience, reducing travel costs in Houzhenzi Town, which has the worst accessibility, by 34%. The expansion model and evaluation method in this study can provide an empirical example for the development of other metropolitan areas and provide a reasonable benchmark and guidance for the development of multi-level rail transit networks in future urban areas.

CLlight: Enhancing representation of multi-agent reinforcement learning with contrastive learning for cooperative traffic signal control

Multi-agent reinforcement learning has shown great potential for coordinating multi-intersection traffic signals due to its powerful adaptive capabilities, treating each intersection as an agent. However, in the real world, different intersections possess differentiating characteristics such as unique vehicle distributions and traffic patterns. Most existing methods directly add neighboring intersection states to local intersections and optimize the cooperative policy network based on synthesized global features. This indirect optimization approach makes it difficult to thoroughly explore the mutual interactions among different intersection agents, preventing agents from truly learning features with cooperative awareness. To resolve these challenges, we introduce contrastive learning as representation task to the multi-intersection traffic signal control approach named CLlight for two-stage policy network updating. In the first stage, we utilize policy-based or actor-critic-based reinforcement learning methods such as A2C, SAC, and PPO to train policy networks with certain representational capabilities. In the second stage, by extracting pre- and post-masked features and reconstructing the post-masked features, the agents are encouraged to learn the similarities and differences between different intersection policies, which in turn enhances the cooperative and individual representation capabilities of the policy network. To the best of our knowledge, this is the first application of contrastive learning in the field of traffic signal control. Experimental results demonstrate, compared to other state-of-the-art traffic signal control methods, superior average travel time and average waiting time performance under various scenarios, tested on synthetic and real-world datasets.

Examining the Influence of Autonomous Vehicle Behaviors on Travel Times and Vehicle Arrivals: A Comparative Study Across Different Simulation Durations on the Kirkuk-Sulaymaniyah Highway

This study delves into the effects of autonomous vehicle behaviors on travel times and vehicle arrivals along the Kirkuk-Sulaymaniyah Highway, employing simulations spanning 3600, 5400, and 7200 seconds. Across varied traffic volumes ranging from 350 to 950 vehicles and autonomous vehicle behaviors categorized as cautious, normal, aggressive, aggressive platoons, and a mix alongside human-driven vehicles, the research unveils significant findings. Results highlight substantial reductions in average travel times and heightened vehicle arrivals among autonomous vehicles, particularly those exhibiting aggressive behaviors, compared to their human-driven counterparts. Across all simulation scenarios, aggressive autonomous vehicles consistently demonstrate superior performance, showcasing potential efficiency gains through aggressive driving algorithms. Furthermore, with increasing traffic volume, the advantages of aggressive autonomous behaviors become more pronounced, suggesting their adaptability to congested conditions. However, safety implications and traffic flow dynamics warrant caution, especially in scenarios with high volumes and aggressive behaviors. These insights underscore the importance of further research and policy considerations to leverage the full potential of autonomous vehicles while ensuring safety and efficiency on highways.

Care seeking during pregnancy: testing the assumptions behind service delivery redesign for maternal and newborn health in rural Kenya

Abstract A health systems reform known as Service Delivery Redesign (SDR) for maternal and newborn health seeks to make high-quality delivery care universal in Kakamega County, in western Kenya, by strengthening hospital-level care and making hospital deliveries the default option for pregnant women. Using a large prospective survey of new mothers in Kakamega County, we examine several key assumptions that underpin the SDR policy’s theory of change. We analyse data on place of delivery, travel time and distance, out-of-pocket spending, and self-reported quality of care for 19 127 women prospectively enrolled during antenatal care (ANC) and surveyed two times after their delivery. We analyze changes in womens’ delivery location preferences in recent years in Kakamega, and over the course of their most recent pregnancy. We also evaluate travel time, out-of-pocket expenditures and patient satisfaction for women who deliver in public hospitals vs primary health centres. We find substantial changes in delivery location at the population level over time and for individual women over the course of pregnancy. Facility delivery has increased from 50.4% in 2010 to 89.5% in 2019; 70% of respondents deliver at a different facility than their reported intention at ANC. Out-of-pocket delivery expenditures are on average 1351 Kenyan shillings (Ksh) in hospitals compared to 964 Ksh in PHC (primary health care)s (P < 0.01). Transport expenditures are 337 Ksh for PHC level deliveries vs 422 Ksh for hospitals (P < 0.01). Self-reported average travel time is 51 min (PHC delivery) vs 47 min (hospital delivery) (P = 0.78). The average distance to a delivery location is 15.1 km for PHC deliveries vs 15.2 km for hospitals (P = 0.99). There were no differences in overall patient-reported quality scores, while some subcomponents of quality favoured hospitals. These findings support several key assumptions of the SDR theory of change in Kakamega County, while also highlighting important challenges that should be addressed to increase the likelihood of successful implementation.

Toward Expanded Access to Cancer Care With Cost Awareness: An Optimization Modeling Analysis of Rwanda.

Cancers are a growing cause of mortality especially in low- and middle-income countries in Africa. Rwanda is no exception. Two cancer centers currently provide care to the public, but there are both political and human interest in expanding access to tertiary cancer care. Improved geographic access could lead to both better patient outcomes and a better understanding of the existing cancer burden across Rwanda. To identify cost-aware ways of expanding geographic access, we adopt an optimization approach and identify expansion plans that minimize the average travel time to a cancer center across the country while remaining under a given monetary budget. Three additional hospitals could reduce average travel times by 40%, with the largest decrease in travel times observed in populations with long travel times. However, such an expansion would require a 50% increase in the number of in-country oncologists. We find that oncologist scarcity, as opposed to monetary constraints, is likely to be a limiting factor for improved access to cancer care. We present an array of expansion plans and suggest that further modeling approaches that incorporate oncologist scarcity can help deliver better policy recommendations.

Evaluation of road roughness and its influence on operating parameters in open-pit mines

Hauling ore and waste rock is a major cost factor in open-pit mines, encompassing both road maintenance and off-road truck operation. Road maintenance costs are directly influenced by traffic volume and the size of trucks using the roads. Off-road truck operating costs are impacted by various factors, including those affecting their movement, such as the direct interaction between tires and the road surface. Tire-road interaction significantly affects rolling resistance. Rougher surfaces with deeper grooves caused by track sinks lead to higher rolling resistance. This study employed laser profiling to characterize the surface roughness of five mine haul roads in two large Brazilian iron ore mines. This study evaluated the impact of these road irregularities on various operating parameters for 240-ton off-highway trucks, including rolling resistance, average travel speed, travel time, productivity, and unit transport cost. The results indicate that a 10 cm increase in road surface roughness can lead to a significant increase in rolling resistance (up to 5%), a substantial decrease in average speed (25%), a notable increase in travel time (26%), a decrease in productivity (19%), and a corresponding increase in unit transport cost (21%).

Analysis of Vehicular Travel Time along the Urban Corridor Using Bluetooth Based Data

There has been a tremendous increase in vehicular growth in the last few decades, which has increased congestion in urban areas. Travel time is one of the measures to understand the congestion levels of the urban corridors. The increased use of Bluetooth devices by vehicular drivers may enable the dynamic data collection of travel time through Bluetooth sensors in urban areas. In this context, the objective of the present study is to evaluate the travel time of vehicles using Bluetooth devices and check the reliability of this Bluetooth technology for traffic studies. To fulfil the stated objective, a Bluetooth application was installed on a smartphone in order to collect the vehicular travel time on a selected corridor. A video camera was setup for traffic data collection at a mid-block section, which is used for the validation of the collected travel time data. The K nearest neighbor (KNN) model was developed with travel time (TT) data. The average TT of cars and other vehicles is estimated and found to be higher in the evening compared to the morning. From the developed KNN model, it is also found that the predicted TT for the next time intervals shows that they follow the trend of the actual travel time data obtained from the Bluetooth app. From the results of this study, it is feasible to estimate the travel time in urban areas using Bluetooth devices, which is useful for engineers and planners for suitable policy decision.

Impact of high-speed railway network on county-level accessibility and economic linkage in Jiangxi Province, China: A spatio-temporal data analysis

Abstract High-speed railway (HSR) networks have profoundly influenced interregional accessibility and economic linkages. This study examined 100 counties (cities) and districts in Jiangxi Province, using spatial and temporal data to measure weighted average travel time, daily accessibility, the accessibility coefficient, and the total and intensity of economic linkages. This study analyzed the impacts of operating a single HSR line versus an entire network. The findings revealed that (1) the HSR network has created a balanced spatial–temporal convergence effect, narrowing the gap between most locations and showing a clear “corridor effect”; (2) the network has strengthened economic connections among counties and has significantly enhanced overall economic output by creating two high-level economic linkage belts, formed around the Shanghai–Kunming and Beijing–Hong Kong HSRs; and (3) the network amplifies the “Siphon effect” and “Matthew effect,” further disadvantaging non-connected regions. Therefore, the province should promote a well-designed, HSR network, enhancing economic exchanges among counties and fostering a high-level economic belt characterized by diversity, complementary advantages, and coordinated development.

Market Penetration Rate Optimization for Mobility Benefits of Connected Vehicles: A Bayesian Optimization Approach

The advancements of connected vehicle (CV) technologies promise significant safety, mobility, and environmental benefits for future transportation systems. These benefits will largely rely on the market penetration rate (MPR) of CVs and connected infrastructure. However, higher MPR is not guaranteed to result in greater benefits in a transportation system in some cases even if we do not consider the deployment cost of CVs. Therefore, understanding the optimal CV MPR to achieve the best system benefits is informative and can provide some guidance for transportation agencies to use appropriate incentives or other policies to potentially affect the speed of CV adoption. Instead of using the traditional incremental method, this paper proposed a simulation-based approach combined with Bayesian optimization to determine the optimal CV MPR that achieves the highest performance benefits for a freeway segment. The proposed methodology is tested in the I-210 E (in California, U.S.) simulation freeway segment built and calibrated in Simulation of Urban Mobility software as a case study. The weighted sum of the average total travel time on the mainline and the average queue length of on-ramps is formulated as the objective function to optimize the CV MPR. Different weight combinations are tested as different scenarios. The optimization results of these scenarios show that, when the weight of total travel time is high, the optimal CV MPR tends to be high. On the contrary, when the weight of queue length increases, higher CV MPRs may not guarantee higher benefits for the traffic system. The globally optimal CV MPR can be as low as 3%. The case study also confirms the effectiveness of optimizing the CV MPR based on microsimulation and Bayesian optimization.

Unveiling referral's impact on asphyxiated newborns: a cross-sectional study

Background: Birth asphyxia also known as perinatal asphyxia, refers to the lack of oxygen during the childbirth. The asphyxiated newborns require a comprehensive care and many special newborns care units have established in Gujarat for the sick newborns. This study mainly focuses on understanding the reasons for referral, the average travel time to tertiary care institutions, and identifying gaps to minimize unnecessary referrals among patients. Methods: The cross-sectional study was conducted over a period of one year, and out-born asphyxiated newborns were selected based on the inclusion criteria. A pretested, semi-structured questionnaire was used to collect the data, which was later analyzed using Microsoft excel and statistical package for the social sciences (SPSS) 26. Results: Out of the 131 newborns, 31% had moderate-severe asphyxia. Approximately 9% of the patients had to travel for more than 2hours to reach the final health facility. Among the patients, risk of mortality was significantly higher among the those who were referred from the private health facility. Lack of availability of bed, ventilator and economic burden were among the common reason for referral. Conclusions: There is a necessity to prevent unnecessary multiple and delayed referrals, and to provide essential care during transport to improve the outcomes.

Reasoning graph-based reinforcement learning to cooperate mixed connected and autonomous traffic at unsignalized intersections

Cooperation at unsignalized intersections in mixed traffic environments, where Connected and Autonomous Vehicles (CAVs) and Manually Driving Vehicles (MVs) coexist, holds promise for improving safety, efficiency, and energy savings. However, the mixed traffic at unsignalized intersections present huge challenges like MVs’ uncertainties, the chain reaction and diverse interactions. Following the thought of the situation-aware cooperation, this paper proposes a Reasoning Graph-based Reinforcement Learning (RGRL) method, which integrates a Graph Neural Network (GNN) based policy and an environment providing mixed traffic with uncertain behaviors. Firstly, it graphicly represents the observed scenario as a situation using the interaction graph with connected but uncertain (bi-directional) edges. The situation reasoning process is formulated as a Reasoning Graph-based Markov Decision Process which infers the vehicle sequence stage by stage so as to sequentially depict the entire situation. Then, a GNN-based policy is constructed, which uses Graph Convolution Networks (GCN) to capture the interrelated chain reactions and Graph Attentions Networks (GAT) to measure the attention of diverse interactions. Furthermore, an environment block is developed for training the policy, which provides trajectory generators for both CAVs and MVs. A reward function that considers social compliance, collision avoidance, efficiency and energy savings is also provided in this block. Finally, three Reinforcement Learning methods, D3QN, PPO and SAC, are implemented for comparative tests to explore the applicability and strength of the framework. The test results demonstrate that the D3QN outperformed the other two methods with a larger converged reward while maintaining a similar converged speed. Compared to multi-agent RL (MARL), the RGRL approach showed superior performance statistically, reduced the number of severe conflicts by 77.78–94.12 %. The RGRL reduced average and maximum travel times by 13.62–16.02 %, and fuel-consumption by 3.38–6.98 % in medium or high Market Penetration Rates (MPRs). Hardware-in-the-loop (HIL) and Vehicle-in-the-loop (VehIL) experiments were conducted to validate the model effectiveness.

Traffic assignment optimization to improve urban air quality with the unified finite-volume physics-informed neural network

Emissions by road transportation constitute the major contributor of air pollutants to threat the public health, especially on mega cities with high-rise buildings and growing population. A practicable and economical remedy is assigning the traffic volume judiciously, termed traffic assignment optimization (TAO). Incorporating the computational fluid dynamics (CFD), this method provides precisely optimized traffic volumes under versatile meteorological conditions. However, the high-dimensional degree of freedoms (DoF) involved in CFD hampers its further application toward the large-scale problem where a massive urban area is considered. With the more complicated urban traffic network, the larger decision variable dimension also impedes the time-efficient acquisition of the optimization results. To alleviate this curse-of-dimensionality, the current work proposes an optimization framework with the surrogate model based on the unified finite-volume physics-informed neural networks (UFV-PINN). The UFV-PINN plays the dual role as the partial differential equation (PDE) solver and the surrogate model for pollutant concentration prediction. It reduces the DoF within the PDE solution process and enables the gradient-based optimization. The current work optimizes the average CO concentration and accumulative travel time in Kowloon peninsula of Hong Kong, using the multiple-gradient descent algorithm (MGDA). This optimization framework is tested with various working conditions. Results indicate that the proposed method attain high accuracy solutions, comparable to the heuristic algorithms. This study is the first attempt to incorporate UFV-PINN into the TAO with air quality consideration. Endowed with the differentiability by UFV-PINN, the framework will facilitate the TAO to provide precise suggestions toward large-scale problems.

Spatiotemporal Accessibility of Rail Transport Systems in the Guangdong–Hong Kong–Macao Greater Bay Area and Its Implications on Economic Equity

Reducing inequality and fostering economic growth is the tenth global goal of the United Nations for sustainable development. Rail transport significantly influences spatial structures, industrial distributions, and is vital for regional economic integration. Despite its importance, the impact of rail transport on economic equity has not been thoroughly explored in current literature. This study aims to fill this gap by evaluating the spatiotemporal characteristics of rail transport accessibility and its implications for economic equity in the Guangdong–Hong Kong–Macao Greater Bay Area. We considered high-speed, intercity, and conventional rail transport and employ three distinct indicators—door-to-door travel time, weighted average travel time, and potential accessibility—to provide a nuanced assessment of accessibility in the region. Each indicator provides a unique perspective on how accessibility affects economic equity, collectively broadening the scope of the analysis. From 1998 to 2020, the evolution of rail transport and its consequent impact on regional economic equity is scrutinized. Advanced econometric methods, namely ordinary least squares, and spatial Durbin models, are combined with the Gini coefficient and Lorenz curve for comprehensive quantitative analysis. This approach highlights the dynamic influence of rail transport development on economic equity, contributing to the sustainable urban development discourse. The results reveal that although rail transport advancements bolster connectivity and economic growth, they also exacerbate regional economic inequality. This study provides valuable insights for urban planning and policymaking by elucidating the complex relationship between rail transport accessibility and economic equity. Our findings underscore the importance of implementing balanced and inclusive transport policies that foster growth and efficiency while mitigating socioeconomic disparities.

Emergency water distribution systems to improve spatial equality and spatial equity in a heterogeneous community with differing mobility characteristics

Our study addresses challenges in emergency water distribution systems by proposing a hybrid method that optimizes points of distribution (PODs) and mobile delivery systems. The goal is to optimally dispense emergency water to disaster-affected populations while enhancing spatial equality and spatial equity. By considering the physiological and socioeconomic status of the disaster-affected population, our hybrid method addresses the needs of a heterogeneous community. The hybrid method consists of two models: The first model seeks to determine the optimal locations of POD for populations who are deemed physiologically able to visit PODs and pick up their emergency water. In this model, socioeconomic status is incorporated to account for different mobility characteristics of these populations. The second model focuses on determining efficient routes for mobile delivery to populations who are more likely to have physiological limitations that interfere with them traveling to PODs and picking up their emergency water. The proposed method is then validated with an application to the Flint, Michigan, water crisis. Our experiments demonstrate that, compared to the actual setup of PODs, our method shows a 69.30 % improvement in objective function value and a 7.05 % reduction in the average travel time for people to reach the PODs. Particularly beneficial for those with the longest travel time to the PODs, the model indicates a significant 25.22 % decrease in travel time, equivalent to 19.49 min. Also, our method suggests the optimal delivery solution involving 20 trucks covering 191.82 km for the target populations. We further conduct a sensitivity analysis to discuss the potential impact of various factors on the operations of the emergency water distribution system. Our results highlight that increasing the number of depots does not necessarily lead to a proportional decrease in vehicle kilometers traveled. We also identify that the most cost-effective vehicle type is a 16-foot truck. These findings provide emergency agencies and policymakers with valuable insights, paving the way for improved guidelines and policies to establish more effective emergency water distribution systems.

Two‐stage algorithm for traffic signal optimization and web‐service system development

AbstractEfficient control of traffic signals for vehicles and pedestrians at intersections is critical for relieving traffic congestion. Considering the unique characteristics of intersections, such as the number of roads, the presence or absence of crosswalks, road geometric shapes, and traffic demand patterns, an appropriate phase sequence and duration for traffic signals must be established at each intersection. This paper proposes a simulation‐based two‐stage algorithm comprising integer‐constrained Adam (ICA) and tabu search (TS) to optimize the phase sequence and duration for arbitrary intersections with arbitrary traffic‐demand patterns. The ICA promptly identifies a promising region in which a global optimal solution is likely to be obtained, whereas TS determines the best solution near the region. The performance of the proposed algorithm that optimizes phase durations with fixed phase sequence is evaluated against several baseline methods using 24 instances across six actual intersections. Experimental results show that the proposed algorithm reduces the average travel time by 20.2% compared with existing traffic signals within a computation time of 4 min, thus providing a near‐optimal solution eight times faster than commonly used population‐based metaheuristics. Furthermore, the algorithm demonstrates robust performance across heterogeneous vehicles and recommends the best phase sequence that effectively alleviates congestion in current traffic signal systems. The optimized phase sequence with best phase durations further reduces the average travel time by approximately 11.3% compared with the existing phase sequence with best phase durations at an actual intersection. To facilitate its widespread use, a free, open web‐service system named “Smart Intersection for Traffic Efficiency” is developed, which enables users to optimize traffic signal systems without requiring optimization background or simulation knowledge.

Short-Term Traffic Flow Prediction Based on Wavelet Analysis and XGBoost

Traffic flow prediction is of great significance for urban planning and alleviating traffic congestion. Due to the randomness and high volatility of urban road network short-term traffic flow, it is difficult for a single model to accurately estimate traffic flow and travel time. In order to obtain more ideal prediction accuracy, a combined prediction model based on wavelet decomposition and reconstruction (WDR) and the extreme gradient boosting (XGBoost) model is developed in this paper. Firstly, the Mallat algorithm is applied to perform multi-scale wavelet decomposition on the average travel time series of the original traffic data, and single branch reconstruction is performed on the components at each scale. Secondly, XGBoost is used to predict each reconstructed single-branch sequence, so as to obtain multiple sub-models, and the Bayesian algorithm is used to optimize the hyperparameters of the sub-models. Finally, the algebraic sum of the predicted values of all sub-models is used to obtain the overall traffic prediction result. To test the performance of the proposed model, actual traffic flow data has been collected from a certain link of the Brooklyn area in New York, USA. The performance of proposed WDR-XGBoost model has been compared with other existing machine learning models, e.g., support vector regression model (SVR) and single XGBoost model. Experimental findings demonstrated that the proposed WDR-XGBoost model performs better on multiple evaluation indicators and has significantly outperformed the other models in terms of accuracy and stability.