Multi-task graph neural network for truck speed prediction under extreme weather conditions

Ramhormozi, Reza Safarzadeh; Mozhdehi, Arash; Kalantari, Saeid; Wang, Yunli; Sun, Sun; Wang, Xin

doi:10.1145/3557915.3561029

Cited by 5 publications

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, these two groups have the best accuracy since the model correctly determines that no overflow will occur in most cases. Note that a common way to deal with imbalanced data is to apply the over-sampling technique such as Smote [34], which improves the classification performance in certain applications [35]. However, including artificial data might affect other metrics.…”

Section: Analysis Of the First Grouping Methodsmentioning

confidence: 99%

Predicting the Overflowing of Urban Personholes Based on Machine Learning Techniques

Chang,

Tseng,

Hsu

2023

Water

View full text Add to dashboard Cite

Urban stormwater drainage systems, which include many personholes to collect and discharge precipitation within a city, are extensively constructed to prevent streets and buildings from flooding. This research intends to build a machine learning model to predict whether a personhole will overflow soon, which is crucial to alleviate the damage caused by floods. To address the challenges posed by many diverse personholes, we proposed segmenting the personholes into several groups and have designed two methods employing different personhole features. The first, the geography-based method, uses the geographical locations of the personholes for the grouping. The second, the hydrology-based method, uses the characteristics that are directly related to the overflowing situation, such as the depth of the personhole, and the average and the maximum water level of the personholes. We also investigated several machine learning techniques, such as the multilayer perceptron (MLP) model and a fine-tuning architecture. The study area was located in the new Taipei city and the experimental results have shown the impressive predictive ability of the proposed approaches. Particularly, by applying the hydrology-based grouping method, and using a hybrid model combining the machine learning model prediction results with heuristic rules, we can obtain the best prediction result, and the accuracy is over 99%. We have also noticed the influence of the activation function used in the neural network and the number of frozen layers in the fine-tuning architecture. Particularly, using the tanh function with one frozen layer is good in some cases. However, since it is not general enough, we suggest the readers perform empirical studies before choosing the best setting in their own environment.

show abstract

Section: Analysis Of the First Grouping Methodsmentioning

confidence: 99%

Predicting the Overflowing of Urban Personholes Based on Machine Learning Techniques

Chang,

Tseng,

Hsu

2023

Water

View full text Add to dashboard Cite

show abstract

“…These issues add to the complexities involved in the trailer allocation and truck routing problem. The advantage of CO on graphs is that this information can be used as node or edge features using methods such as Spatio-Temporal Graph Convolutional Networks (ST-GCN) (Yu et al, 2017) for considering spatial and temporal information on the traffic network or the Multi-Task Context Based Gated Recurrent Unit Graph Convolutional Network (MT-C2G) (Ramhormozi et al, 2022) for incorporating different contextual information such as weather conditions, time of the week, holidays, traffic speed and volume in the embedding of the graph.…”

Section: Con Clus Ionmentioning

confidence: 99%

Trailer allocation and truck routing using bipartite graph assignment and deep reinforcement learning

Kalantari

Ramhormozi

Wang

et al. 2023

Transactions in GIS

Self Cite

View full text Add to dashboard Cite

Trailer allocation and truck routing are critical components of truck transportation management. However, in real-world applications, inter-influence between selecting the best trailers and trucks, strict fulfillment of Pickup or Delivery (PD) orders, and the size of the fleet are some of the challenges that need to be dealt with in a large truck company. In addition, trailer allocation and truck routing problems are considered to be NP-hard combinatorial optimization (CO) problems. Therefore, we use deep reinforcement learning (DRL), which has the capability of solving routing problems with a single set of hyperparameters. This is significant progress toward finding strong heuristics for a special case of the trailer allocation to customers and truck routing problem presented in this article. Given a set of trailers, trucks, customers, and orders we propose a novel two-phase framework based on Bipartite Graph Assignment (BGA) and attention-based DRL to minimize the total traveling distance traveled from trucks to trailers and then to customers. The BGA heuristic finds the minimum traveling distance from the trailers to the customers based on the edge information and the encoder-decoder helps DRL to get useful node and graph feature representations and trains the model to find the proper solutions | 997 KALANTARI et al.

show abstract

Machine learning with SHapley additive exPlanations for evaluating mine truck productivity under real-site weather conditions at varying temporal resolutions

Fan,

Arachchilage,

Zhang

et al. 2024

International Journal of Mining, Reclamation and Environment

View full text Add to dashboard Cite

Multi-task graph neural network for truck speed prediction under extreme weather conditions

Cited by 5 publications

References 21 publications

Predicting the Overflowing of Urban Personholes Based on Machine Learning Techniques

Predicting the Overflowing of Urban Personholes Based on Machine Learning Techniques

Trailer allocation and truck routing using bipartite graph assignment and deep reinforcement learning

Machine learning with SHapley additive exPlanations for evaluating mine truck productivity under real-site weather conditions at varying temporal resolutions

Contact Info

Product

Resources

About