Lina Yu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Lina Yu has a total of 48 co-authors. Here are the top co-authors by number of shared publications:
- Zhuang Wu
School of Management and Engineering Capital University of Economics and Business 100070 Beijing China
5 shared publications - Jing Gao
Capital University of Economics and Business, Beijing, China
4 shared publications Qin Luo
3 shared publicationsQunzhi Zhou
3 shared publications- Huasheng Yang
Department of Industrial Engineering, Tsinghua University, Beijing, China
2 shared publications Huawei Zheng
2 shared publications
Latest Publications
- Open Access
- 15 September 2026
- Published by IOS Press
- p. 927-933
Single-Agent Reinforcement Learning Model for Regional Adaptive Traffic Signal Control
Several studies have employed reinforcement learning (RL) to address the challenges of regional adaptive traffic signal control (ATSC) and achieved promising results. In this field, existing research predominantly adopts multi-agent frameworks. However, the adoption of multi-agent frameworks presents challenges for scalability. Instead, the Traffic signal control (TSC) problem necessitates a single-agent framework. TSC inherently relies on centralized management by a single control center, which can monitor traffic conditions across all roads in the study area and coordinate the control of all intersections. This work proposes a single-agent RL-based regional ATSC model compatible with probe vehicle technology. Key components of the RL design include state, action, and reward function definitions. To facilitate learning and manage congestion, both state and reward functions are defined based on queue length, with action designed to regulate queue dynamics. The queue length definition used in this study differs slightly from conventional definitions but is closely correlated with congestion states. More importantly, it allows for reliable estimation using link travel time data from probe vehicles. With probe vehicle data already covering most urban roads, this feature enhances the proposed method’s potential for widespread deployment. The method was comprehensively evaluated using the SUMO simulation platform. Experimental results demonstrate that the proposed model significantly reduces queue variability and decreases queue lengths to 63% of the baseline level, confirming that it effectively mitigates large-scale regional congestion levels via coordinated multi-intersection control.
- 15 September 2026
- Published by IOS Press
- p. 927-933
Open AccessSingle-Agent Reinforcement Learning Model for Regional Adaptive Traffic Signal Control
Several studies have employed reinforcement learning (RL) to address the challenges of regional adaptive traffic signal control (ATSC) and achieved promising results. In this field, existing research predominantly adopts multi-agent frameworks. However, the adoption of multi-agent frameworks presents challenges for scalability. Instead, the Traffic signal control (TSC) problem necessitates a single-agent framework. TSC inherently relies on centralized management by a single control center, which can monitor traffic conditions across all roads in the study area and coordinate the control of all intersections. This work proposes a single-agent RL-based regional ATSC model compatible with probe vehicle technology. Key components of the RL design include state, action, and reward function definitions. To facilitate learning and manage congestion, both state and reward functions are defined based on queue length, with action designed to regulate queue dynamics. The queue length definition used in this study differs slightly from conventional definitions but is closely correlated with congestion states. More importantly, it allows for reliable estimation using link travel time data from probe vehicles. With probe vehicle data already covering most urban roads, this feature enhances the proposed method’s potential for widespread deployment. The method was comprehensively evaluated using the SUMO simulation platform. Experimental results demonstrate that the proposed model significantly reduces queue variability and decreases queue lengths to 63% of the baseline level, confirming that it effectively mitigates large-scale regional congestion levels via coordinated multi-intersection control.
- Open Access
- 8 July 2026
- Published by MDPI AG in Sustainability
- Vol. 18 (14) , 6970
Typhoon Disaster Chain Evolution Modelling in the Guangdong–Hong Kong–Macao Greater Bay Area Based on GERT Stochastic Networks: A Case Study of Typhoon Mangkhut
Typhoon disaster chains—cascading sequences of hazards triggered by tropical cyclones—pose significant risks to highly urbanized coastal regions. However, quantitative modelling of their stochastic evolution, including both propagation probabilities and temporal dynamics, remains limited. This study develops a typhoon disaster chain evolution model by integrating the FBIREC (Factor-Bearing body- Incident-Response-Environment-Consequence) scenario representation framework with a Graphical Evaluation and Review Technique (GERT) stochastic network. The model is applied to the Guangdong–Hong Kong–Macao Greater Bay Area using Typhoon Mangkhut (2018) as a case study. Ten cascading event sequences are identified, and their transmission probabilities and expected durations are calculated. The model reproduces the observed event probabilities for Typhoon Mangkhut with high fidelity, with a mean absolute error of 4.31% and a coefficient of determination R2 = 0.816. Power facility damage was identified as one of the most influential infrastructure nodes, exhibiting high propagation potential to multiple downstream service systems. The disaster chain unfolds across two temporal scales: short-term risks (0–48 h) and long-term risks (>48 h), providing quantifiable targets for phased emergency response planning. The proposed framework provides a methodological reference that can be adapted to other typhoon-prone regions following case-specific re-parameterization.
- 8 July 2026
- Published by MDPI AG in Sustainability
- Vol. 18 (14) , 6970
Open AccessTyphoon Disaster Chain Evolution Modelling in the Guangdong–Hong Kong–Macao Greater Bay Area Based on GERT Stochastic Networks: A Case Study of Typhoon Mangkhut
Typhoon disaster chains—cascading sequences of hazards triggered by tropical cyclones—pose significant risks to highly urbanized coastal regions. However, quantitative modelling of their stochastic evolution, including both propagation probabilities and temporal dynamics, remains limited. This study develops a typhoon disaster chain evolution model by integrating the FBIREC (Factor-Bearing body- Incident-Response-Environment-Consequence) scenario representation framework with a Graphical Evaluation and Review Technique (GERT) stochastic network. The model is applied to the Guangdong–Hong Kong–Macao Greater Bay Area using Typhoon Mangkhut (2018) as a case study. Ten cascading event sequences are identified, and their transmission probabilities and expected durations are calculated. The model reproduces the observed event probabilities for Typhoon Mangkhut with high fidelity, with a mean absolute error of 4.31% and a coefficient of determination R2 = 0.816. Power facility damage was identified as one of the most influential infrastructure nodes, exhibiting high propagation potential to multiple downstream service systems. The disaster chain unfolds across two temporal scales: short-term risks (0–48 h) and long-term risks (>48 h), providing quantifiable targets for phased emergency response planning. The proposed framework provides a methodological reference that can be adapted to other typhoon-prone regions following case-specific re-parameterization.
- 1 April 2026
- Published by Springer Nature
- p. 246-255
Retrieval Time Optimization for Puzzle-Based Storage Systems Considering Direction Switching
- 1 April 2026
- Published by Springer Nature
- p. 246-255
Retrieval Time Optimization for Puzzle-Based Storage Systems Considering Direction Switching
- 1 April 2026
- Published by Springer Nature
- p. 114-122
Traffic Signal-Constrained Improved A-Star Algorithm for Multi-Station Collaborative Fire Emergency Dispatch
- 1 April 2026
- Published by Springer Nature
- p. 114-122
Traffic Signal-Constrained Improved A-Star Algorithm for Multi-Station Collaborative Fire Emergency Dispatch
- 5 March 2026
- Published by Elsevier in Computers & Industrial Engineering
- Vol. 216
A traveling salesman problem with drone stations and speed-optimized drones
- 5 March 2026
- Published by Elsevier in Computers & Industrial Engineering
- Vol. 216
A traveling salesman problem with drone stations and speed-optimized drones
- 10 January 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 271
A reliability-oriented conflict robust optimization model for berth allocation under operational disturbances
- 10 January 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 271
A reliability-oriented conflict robust optimization model for berth allocation under operational disturbances
- 10 October 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 186
A novel spatio-temporal adaptive network considering time-delay effect for traffic flow forecasting
- 10 October 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 186
A novel spatio-temporal adaptive network considering time-delay effect for traffic flow forecasting
- Open Access
- 26 August 2025
- Published by IOS Press
- p. 88-98
Research on Picking Algorithm of Puzzle-Based Storage System Based on Reinforcement Learning
The puzzle-based storage system (PBS) is an emerging intensive storage system that eliminates all aisles in the aisle-based shuttle storage system, significantly enhancing the storage efficiency of the warehouse. This paper aims to explore and apply reinforcement learning algorithms to PBS, offering an innovative solution. Firstly, the principles of the proposed algorithm, including the learning agent and actions, the environment and states, and reward functions, are presented in detail. Then the feasibility of Deep Q-Network (DQN) and Proximal Policy Optimization (PPO) algorithms for optimizing pick-up efficiency are validated. Finally, the experimental results show that the proposed reinforcement learning approach with satisfactory performance for smaller systems, but varying convergence for larger ones, highlighting the importance of training iterations for model stability and performance.
- 26 August 2025
- Published by IOS Press
- p. 88-98
Open AccessResearch on Picking Algorithm of Puzzle-Based Storage System Based on Reinforcement Learning
The puzzle-based storage system (PBS) is an emerging intensive storage system that eliminates all aisles in the aisle-based shuttle storage system, significantly enhancing the storage efficiency of the warehouse. This paper aims to explore and apply reinforcement learning algorithms to PBS, offering an innovative solution. Firstly, the principles of the proposed algorithm, including the learning agent and actions, the environment and states, and reward functions, are presented in detail. Then the feasibility of Deep Q-Network (DQN) and Proximal Policy Optimization (PPO) algorithms for optimizing pick-up efficiency are validated. Finally, the experimental results show that the proposed reinforcement learning approach with satisfactory performance for smaller systems, but varying convergence for larger ones, highlighting the importance of training iterations for model stability and performance.
- 23 July 2025
- Published by Elsevier in Neural Networks
- Vol. 192, 107905
Reinforcement learning with temporal and variable dependency-aware transformer for stock trading optimization
- 23 July 2025
- Published by Elsevier in Neural Networks
- Vol. 192, 107905
Reinforcement learning with temporal and variable dependency-aware transformer for stock trading optimization
- Open Access
- 24 April 2025
- Published by Acadlore Publishing Services Limited in Acadlore Transactions on AI and Machine Learning
- Vol. 4 (2) , 109-123
Customer Churn Prediction in the Banking Sector Using Sentence Transformers and a Stacking Ensemble Framework
As market saturation and competitive pressure intensify within the banking sector, the mitigation of customer churn has emerged as a critical concern. Given that the cost of acquiring new clients substantially exceeds that of retaining existing ones, the development of highly accurate churn prediction models has become imperative. In this study, a hybrid customer churn prediction model was developed by integrating Sentence Transformers with a stacking ensemble learning architecture. Customer behavioral data containing textual content was transformed into dense vector representations through the use of Sentence Transformers, thereby capturing contextual and semantic nuances. These embeddings were combined with normalized structured features. To enhance predictive performance, a stacking ensemble method was employed to integrate the outputs of multiple base models, including random forest, Gradient Boosting Tree (GBT), and Support Vector Machine (SVM). Experimental evaluation was conducted on real-world banking data, and the proposed model demonstrated superior performance relative to conventional baseline approaches, achieving notable improvements in both accuracy and the area under the curve (AUC). Furthermore, the analysis of model outputs revealed several salient predictors of customer attrition, such as anomalous transaction behavior, prolonged inactivity, and indicators of dissatisfaction with customer service. These insights are expected to inform the development of targeted intervention strategies aimed at strengthening customer retention, improving satisfaction, and fostering long-term institutional growth and stability.
- 24 April 2025
- Published by Acadlore Publishing Services Limited in Acadlore Transactions on AI and Machine Learning
- Vol. 4 (2) , 109-123
Open AccessCustomer Churn Prediction in the Banking Sector Using Sentence Transformers and a Stacking Ensemble Framework
As market saturation and competitive pressure intensify within the banking sector, the mitigation of customer churn has emerged as a critical concern. Given that the cost of acquiring new clients substantially exceeds that of retaining existing ones, the development of highly accurate churn prediction models has become imperative. In this study, a hybrid customer churn prediction model was developed by integrating Sentence Transformers with a stacking ensemble learning architecture. Customer behavioral data containing textual content was transformed into dense vector representations through the use of Sentence Transformers, thereby capturing contextual and semantic nuances. These embeddings were combined with normalized structured features. To enhance predictive performance, a stacking ensemble method was employed to integrate the outputs of multiple base models, including random forest, Gradient Boosting Tree (GBT), and Support Vector Machine (SVM). Experimental evaluation was conducted on real-world banking data, and the proposed model demonstrated superior performance relative to conventional baseline approaches, achieving notable improvements in both accuracy and the area under the curve (AUC). Furthermore, the analysis of model outputs revealed several salient predictors of customer attrition, such as anomalous transaction behavior, prolonged inactivity, and indicators of dissatisfaction with customer service. These insights are expected to inform the development of targeted intervention strategies aimed at strengthening customer retention, improving satisfaction, and fostering long-term institutional growth and stability.