Practical End-to-End Repositioning Algorithm for Managing Bike-Sharing System

Yoshida, Akihiro; Yatsushiro, Yosuke; Hata, Nozomi; Higurashi, Tatsuru; Tateiwa, Nariaki; Wakamatsu, Takashi; Nagamatsu, Kenichi

doi:10.1109/bigdata47090.2019.9005986

Cited by 10 publications

(7 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An analogous problem that has received much more attention in the past is the prediction of the availability of bicycles at public bicycle sharing stations [26][27][28][29][30][31][32][33][34][35]. The problems are comparable since there are a number of spots that can either be occupied or not and it is relevant to determine when stations are out of bicycles to rent.…”

Section: Literature Reviewmentioning

confidence: 99%

“…In the context of this paper, the aggregation unit is the individual CS, with the corresponding high noise making such tools unsuitable. A selection of employed machine learning models includes random forest models [26], deep learning models [27], k-means clustering [31,32], or, less frequently, models such as the averaged one dependence estimators with subsumption resolution model [33]. Most of these models are in fact quite useful for our purpose as well and have been investigated.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Most of these models are in fact quite useful for our purpose as well and have been investigated. Deep learning, such as done by Yoshida et al [27], however, requires that more input variables are available per decision than is possible in this paper. In other words, a CS history is not relevant for future events apart from the average weekly usage.…”

Section: Literature Reviewmentioning

confidence: 99%

See 2 more Smart Citations

Predicting Electric Vehicle Charging Station Availability Using Ensemble Machine Learning

2021

View full text Add to dashboard Cite

Electric vehicles may reduce greenhouse gas emissions from individual mobility. Due to the long charging times, accurate planning is necessary, for which the availability of charging infrastructure must be known. In this paper, we show how the occupation status of charging infrastructure can be predicted for the next day using machine learning models— Gradient Boosting Classifier and Random Forest Classifier. Since both are ensemble models, binary training data (occupied vs. available) can be used to provide a certainty measure for predictions. The prediction may be used to adapt prices in a high-load scenario, predict grid stress, or forecast available power for smart or bidirectional charging. The models were chosen based on an evaluation of 13 different, typically used machine learning models. We show that it is necessary to know past charging station usage in order to predict future usage. Other features such as traffic density or weather have a limited effect. We show that a Gradient Boosting Classifier achieves 94.8% accuracy and a Matthews correlation coefficient of 0.838, making ensemble models a suitable tool. We further demonstrate how a model trained on binary data can perform non-binary predictions to give predictions in the categories “low likelihood” to “high likelihood”.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

See 1 more Smart Citation

Predicting Electric Vehicle Charging Station Availability Using Ensemble Machine Learning

2021

View full text Add to dashboard Cite

show abstract

“…However, when used to predict bike-sharing at the station level, CNN can only reflect inter-station relationship by geographical distance [39]. Some researchers attempted to apply deep learning architecture to graph data structure [39][40][41][42]. Taking bike stations as nodes, the bike-sharing network can be represented in a graph.…”

Section: A Short-term Bike-sharing Demand Predictionmentioning

confidence: 99%

“…Kim et al [39] constructed a GCN prediction model to predict hourly bike-sharing demand at the station level by incorporating spatial characteristics, temporal patterns, and global variables (weather and weekday/weekend). Yoshida et al [40] proposed a relational graph convolutional networkbased method to predict the demand at the station level. Guo et al [41] built a spatial-temporal graph neural network (ST-GNN) to model and predicted citywide bike-sharing demand.…”

Section: A Short-term Bike-sharing Demand Predictionmentioning

confidence: 99%

Station Importance Evaluation in Dynamic Bike-Sharing Rebalancing Optimization Using an Entropy-Based TOPSIS Approach

Jin

2021

IEEE Access

View full text Add to dashboard Cite

As an eco-friendly travel mode, bike-sharing has prevailed around the world. However, the systems are imbalanced due to the asymmetric spatial and temporal distribution of user demand. Station prioritization strategies are needed to rebalance more shared bikes for more important stations. This paper proposes an evaluation method of station importance in dynamic bike-sharing rebalancing. Firstly, a shortterm demand prediction model is applied to capture the temporal and spatial characteristics of bike-sharing trip data and predict bike-sharing demand at the station level. Based on the prediction results, the method of determining rebalancing quantity is proposed with consideration of bike-sharing usage throughout the rebalancing period. Then, three criteria are employed to evaluate the importance of bike-sharing stations, including rebalancing quantity, closeness to inventory threshold, and distance from the key station. An entropy-based Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) approach is proposed to weigh different criteria and evaluate station importance. Furthermore, the experiments on bikesharing data from Nanjing City demonstrate the effectiveness of the proposed methods. This research is helpful for operators and managers to dynamically rebalance shared bikes with high efficiency and improve the service quality of bike-sharing systems.

show abstract

An AI Approach to Rebalance Bike-Sharing Systems with Adaptive User Incentive

Duan

2022

Internet of Things

View full text Add to dashboard Cite

Practical End-to-End Repositioning Algorithm for Managing Bike-Sharing System

Cited by 10 publications

References 16 publications

Predicting Electric Vehicle Charging Station Availability Using Ensemble Machine Learning

Predicting Electric Vehicle Charging Station Availability Using Ensemble Machine Learning

Station Importance Evaluation in Dynamic Bike-Sharing Rebalancing Optimization Using an Entropy-Based TOPSIS Approach

An AI Approach to Rebalance Bike-Sharing Systems with Adaptive User Incentive

Contact Info

Product

Resources

About