Low-carbon benefits of aircraft adopting continuous descent operations

Applied Energy, Journal Year: 2025, Volume and Issue: 383, P. 125390 - 125390

Published: Jan. 21, 2025

Language: Английский

---

Data-driven Analysis of Taxi and Ride-hailing Services: Case Study in Chengdu, China

Computer and decision making., Journal Year: 2025, Volume and Issue: 2, P. 357 - 373

Published: Jan. 3, 2025

Both taxi and Internet-based ride-hailing services are important public transportation options. While these provide a similar functionality, their quantitative patterns difficult to describe analyze without big data approach. In this study, we present data-driven analysis for with case study in Chengdu, China. GPS dataset an trip provided by Didi Chuxing, the largest network company China, used study. Our is based on four aspects, i.e., temporal patterns, spatial spatio-temporal traveling distance patterns. It found that both exhibit densification power law space-time graph model. The service has preference more concentrated pickup dropoff hotspots trips longer distance. observed results indicate partially plays role of taxis, but it also its own characteristics as new transport option. findings would be helpful government better regulate operation services.

Language: Английский

---

Graph Convolutional Networks with Multi-Scale Dynamics for Traffic Speed Forecasting

Applied Soft Computing, Journal Year: 2025, Volume and Issue: unknown, P. 112966 - 112966

Published: March 1, 2025

Language: Английский

---

Origin–destination prediction via knowledge‐enhanced hybrid learning

Computer-Aided Civil and Infrastructure Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: March 18, 2025

Abstract This paper proposes a novel origin–destination (OD) prediction (ODP) model, namely, knowledge‐enhanced hybrid spatial–temporal graph neural networks (KE‐H‐GNN). KE‐H‐GNN integrates deep learning predictive model with traffic engineering domain knowledge and multi‐linear regression (MLR) module for incorporating external factors. Leveraging insights from the gravity we propose two meaningful region partitioning strategies reducing data dimension: election districts K‐means clustering. The aggregated OD matrices inputs are processed using an long short‐term memory network to capture temporal correlations multi‐graph input convolutional spatial correlations. also employs global–local attention module, inspired by flow theory, nonlinear features. Finally, MLR was designed quantify relationship between Experiments on real‐world datasets New York Tokyo demonstrate that outperforms all baseline models while maintaining interpretability. Additionally, outperformed concatenation method integrating factors, regarding both performance transparency. Moreover, district‐based approach proved more effective simpler practical applications. proposed offers interpretable solution ODP can be practically applied in scenarios.

Language: Английский

---

Multivariate N-bEATS Algorithm for Vehicular Traffic Flow Prediction Analysis For Indian Road Scenario

International Journal of Intelligent Transportation Systems Research, Journal Year: 2025, Volume and Issue: unknown

Published: May 13, 2025

Language: Английский

---

Graph Neural Networks for Routing Optimization: Challenges and Opportunities

Sustainability, Journal Year: 2024, Volume and Issue: 16(21), P. 9239 - 9239

Published: Oct. 24, 2024

In this paper, we explore the emerging role of graph neural networks (GNNs) in optimizing routing for next-generation communication networks. Traditional protocols, such as OSPF or Dijkstra algorithm, often fall short handling complexity, scalability, and dynamic nature modern network environments, including unmanned aerial vehicle (UAV), satellite, 5G By leveraging their ability to model topologies learn from complex interdependencies between nodes links, GNNs offer a promising solution distributed scalable optimization. This paper provides comprehensive review latest research on GNN-based methods, categorizing them into supervised learning modeling, optimization, reinforcement tasks. We also present detailed analysis existing datasets, tools, benchmarking practices. Key challenges related real-world deployment, explainability, security are discussed, alongside future directions that involve federated learning, self-supervised online techniques further enhance GNN applicability. study serves first survey aiming inspire practical applications

Language: Английский

---

Improved air traffic flow prediction in terminal areas using a multimodal spatial–temporal network for weather-aware (MST-WA) model

Advanced Engineering Informatics, Journal Year: 2024, Volume and Issue: 62, P. 102935 - 102935

Published: Oct. 1, 2024

Language: Английский

---

An Efficient Algorithm for Extracting Railway Tracks Based on Spatial-Channel Graph Convolutional Network and Deep Neural Residual Network

ISPRS International Journal of Geo-Information, Journal Year: 2024, Volume and Issue: 13(9), P. 309 - 309

Published: Aug. 29, 2024

The accurate detection of railway tracks is essential for ensuring the safe operation railways. This study introduces an innovative algorithm that utilizes a graph convolutional network (GCN) and deep neural residual to enhance feature extraction from high-resolution aerial imagery. traditional encoder–decoder architecture expanded with GCN, which improves neighborhood definitions enables long-range information exchange in single layer. As result, complex track features contextual are captured more effectively. network, incorporates depthwise separable convolution inverted bottleneck design, representation long-distance positional addresses occlusion caused by train carriages. scSE attention mechanism reduces noise optimizes representation. was trained tested on custom Massachusetts datasets, demonstrating 89.79% recall rate. 3.17% improvement over original U-Net model, indicating excellent performance segmentation. These findings suggest proposed not only excels segmentation but also offers significant competitive advantages performance.

Language: Английский

---

Dynamic data reconciliation for enhancing the prediction performance of long short-term memory network

Measurement Science and Technology, Journal Year: 2024, Volume and Issue: 35(11), P. 116147 - 116147

Published: Aug. 19, 2024

Abstract In modern process industries, long short-term memory (LSTM) network is widely used for data-driven modeling. Constrained by measuring instruments and environments, the measured datasets are generally with Gaussian/non-Gaussian distributed measurement noise. The noisy will impact modeling accuracy of LSTM decrease prediction performance it. Aiming at addressing impairment under distribution, this study introduces dynamic data reconciliation (DDR) both into training test. Results show that DDR improves not only quality based on outputs via Bayesian formula in model step, but also offline information test outputs. implementation scheme Gaussian non-Gaussian noise purposely designed. effectiveness verified a numerical example case involving set shared wind power datasets.

Language: Английский

---

PSTCGCN: Principal spatio-temporal causal graph convolutional network for traffic flow prediction

Neural Computing and Applications, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 26, 2024

Language: Английский

---