Jingxu Chen’s research while affiliated with Southeast University and other places

div class="section abstract"> Due to the crucial impact on flight scheduling, airline planning, and airport operations, flight departure delay prediction has emerged as a severe and prominent issue within the realm of smart aviation systems. Accurately predicting flight departure delay durations constitutes a crucial aspect of smart aviation management. Such predictive capability empowers aviation authorities and airport regulators to implement optimized air traffic control strategies, mitigating delays and elevating airport operational efficiency, while enhancing the satisfaction of travelers. The methodology employed in flight delay prediction has undergone substantial evolution in recent years, progressing from rudimentary statistical models to more sophisticated and intricate machine learning models. In this study, we introduce a novel machine learning model enriched with network features and grid search-based parameter selection for advanced predictive analytics of flight departure delays. This model integrates air traffic network feature extraction, feature selection, and machine learning-based prediction. Specifically, we leverage complex network theory to extract both node-level and edge-level features from the air traffic network. Subsequently, the XGBoost algorithm is employed for feature selection and delay prediction, capitalizing on its flexibility and robust performance. A case study utilizing a high-dimensional flight dataset from the U.S. Bureau of Transportation Statistics (BTS) was conducted to assess the model’s effectiveness. The experimental results and the visualization results demonstrate that the proposed framework surpasses several benchmark models, achieving an average delay prediction accuracy with a deviation of about 3.7 minutes. This framework exhibits strong potential for addressing high-dimensional, large-scale predictive challenges in flight delay management while maintaining superior accuracy. </div

Enabled by recent technological advances and the substantial growth of the sharing economy, electric bike-sharing (EBS) has experienced rapid growth in medium-sized Chinese cities, yet its impact on for-hire vehicle (FHV) services remains insufficiently studied. Using a six-month longitudinal dataset from Yancheng, a representative medium-sized city in China, we employ an instrumental variable method to address potential endogeneity and provide quantitative empirical analysis. The analysis identifies a significant substitution effect, where a 1% increase in EBS trips corresponds to a 0.810% decline in FHV ridership. Through heterogeneity analyses, this study reveals that the substitutive effect of EBS is stronger in central downtown, which has denser infrastructure, while its impact diminishes in peripheral districts. Furthermore, unfavorable weather conditions mitigate the substitutive effect, as users increasingly rely on FHVs for their reliability and comfort during unfavorable conditions. The findings of this study highlight the necessity of integrating EBS into the electrified shared mobility ecosystem in a balanced manner to prevent disruptions to the existing transportation network and provide valuable guidance for sustainable and stable transportation planning in medium-sized cities and similar urban contexts.

Knowledge of vacant taxis’ passenger-searching behavior is of great social and economic interest to multiple applications, particularly in planning and operating on-demand mobility services. The inherent stochasticity and dynamic nature in both passenger demand and drivers’ decision-making pose challenges in capturing vacant taxis movements. This study proposes a novel deep clustering framework to comprehensively uncover passenger-searching strategies from extensive trajectory data. Specifically, multiple features from vacant searching trips are extracted and analogously defined as multi-channel images, where each channel corresponds to a specific feature. A novel deep image clustering approach is then proposed, integrating a feature representation module utilizing convolutional neural networks (ConvNets), a self-expression module for affinity learning, a spectral clustering module, and a classification module for self-supervision. An effective training procedure is also presented for the proposed deep clustering framework. Experiments demonstrate the effectiveness of the proposed approach against benchmark methods. Based on the clustering results, common and specific passenger searching strategies are further revealed. Specifically, our findings highlight the importance of individual’s contextual experience in explaining searching behavior and operational efficiency. Moreover, drivers cruising without clear searching strategy often exhibit lower performance, and some drivers may gamble with peers to increase their chances of picking up passengers. These results deliver important justifications for future studies and provide managerial implications to improve on-demand mobility.

In transportation infrastructure systems, feature images and spatial characteristics are generally utilized as complementary elements derived from point clouds for road edge extraction, but the involvement of one or more hyperparameters in each makes the extraction complicated. This study proposes an autotuning hybrid method with Bayesian optimization for road edge extraction in highway systems. The hybrid method combines the strengths of 2D feature images and 3D spatial characteristics while also automatically tuning the hyperparameter combination using Bayesian optimization. The hyperparameters encompass high and low pixel gradient thresholds, neighborhood radius, and normal vector threshold. Later, the point cloud dataset of national highways in Henan Province, China, is taken as the case study to evaluate the performance of the proposed method against three benchmark methods in two typical road scenarios: straight and curved edges. Experimental results show that the proposed method outperforms the benchmarks in detection quality and accuracy. It can serve as a decision-making tool to complement traditional manual road surveying, enabling efficient and automated road edge extraction in highway systems.

In many cities, bike sharing systems, including station-based bike sharing (SBBS) and dockless bike sharing (DBS), are gaining popularity rapidly. Bike rebalancing is one of the most expensive aspects of bike sharing operations, and it takes several hours. In terms of reducing the inefficiencies of frequent short-term bike reba-lancing, whether bike distribution achieves long-term self-balance for one day or even longer periods is a critical issue that has received insufficient attention. This paper aims to provide insights into long-term facility planning by investigating the self-balancing phenomenon of shared bikes. It is evaluated using daily stability analyses from the DBS case in Nanjing, China, and the SBBS case in New York, USA. DBS virtual stations were identified throughout the city, and (virtual) stations can be classified into four categories using various clustering methods. The findings demonstrate that 72% of DBS virtual stations and 81% of SBBS stations can achieve bike self-balancing, with only a few (virtual) stations failing to do so. In terms of non-self-balancing stations, bike-increasing stations are primarily located in the city center, whereas bike-fluctuating stations are primarily found near metro lines. This research can assist bike sharing companies with their daily operations and contribute to government management.

Expressway systems play a vital role in facilitating intercity travels for both passengers and freights, which are also a significant source of vehicle emissions within the transportation sector. This study investigates vehicle emissions from expressway systems using the COPERT model to develop multi-year emission inventories for different pollutants, covering the past and future trends from 2005 to 2030. Thereinto, an integrated SARIMA-SVR method is designed to portray the temporal variation of vehicle population, and the possible future trends of expressway vehicle emissions are predicted through policy scenario analysis. The Jiang–Zhe–Hu Region of China is taken as the case study to analyze emission control in expressway systems. The results indicate that (1) carbon monoxide (CO) and volatile organic compounds (VOCs) present a general upward trend primarily originating from passenger vehicles, while nitrogen oxides (NOx) and inhalable particles (PM) display a slowing upward trend with fluctuations mainly sourcing from freight vehicles; (2) vehicle population constraint is an effective emission control policy, but upgrading the medium- and long-haul transportation structure is necessary to meet the continuous growth of intercity trips. Expressway vehicle emission reduction effectiveness can be further enhanced by curtailing the update frequency of emission standards, along with the scrapping of high-emission vehicles.