A5100385816- Vehicle Routing Optimization Methods
- Supply Chain and Inventory Management
- Metaheuristic Optimization Algorithms Research
- Optimization and Packing Problems
- Human Pose and Action Recognition
- Topic Modeling
- Scheduling and Optimization Algorithms
- Electrocatalysts for Energy Conversion
- Advanced Neural Network Applications
- Stock Market Forecasting Methods
- Fuel Cells and Related Materials
- Neural Networks and Reservoir Computing
- Gait Recognition and Analysis
- Catalytic Processes in Materials Science
- Structural Health Monitoring Techniques
- E-commerce and Technology Innovations
- Remote Sensing and Land Use
- Sustainable Supply Chain Management
- Blind Source Separation Techniques
- Chinese history and philosophy
- Infrastructure Maintenance and Monitoring
- Catalysts for Methane Reforming
- Big Data and Digital Economy
- Hong Kong and Taiwan Politics
- Surface Treatment and Coatings
Tianjin University
2022-2025
Harbin Institute of Technology
2017-2025
Yancheng Teachers University
2025
Southwest Petroleum University
2024
Shanghai Maritime University
2019-2024
Guangxi University
2024
University of California, Irvine
2022-2023
Shaanxi Normal University
2023
City University of Hong Kong
2023
Delta International (Georgia)
2023
Uncertain Capacitated Arc Routing Problem (UCARP) is a variant of the well-known CARP. It considers variety stochastic factors to reflect reality where exact information such as actual task demand and accessibilities edges are unknown in advance. Existing works focus on obtaining robust solution beforehand. However, it also important design effective heuristics adjust real time. In this paper, we develop new Genetic Programming-based Hyper-Heuristic (GPHH) for automated heuristic UCARP. A...
The uncertain capacitated arc routing problem is of great significance for its wide applications in the real world. In problem, variables such as task demands and travel costs are realised time. This may cause predefined solution to become ineffective and/or infeasible. There two main challenges solving this problem. One obtain a high-quality robust baseline sequence, other design an effective recourse policy adjust sequence when it becomes infeasible during execution. Existing studies...
Abstract In recent years, human pose estimation has been widely studied as a branch task of computer vision. Human plays an important role in the development medicine, fitness, virtual reality, and other fields. Early technology used traditional manual modeling methods. Recently, developed rapidly using deep learning. This study not only reviews basic research but also summarizes latest cutting-edge technologies. addition to systematically summarizing technology, this article extends...
With the advent of Transformer, attention mechanism has been applied to Large Language Model (LLM), evolving from initial single- modal large models today's multi-modal models. This greatly propelled development Artificial Intelligence (AI) and ushered humans into era Single-modal can be broadly categorized three types based on their application domains: Text LLM for Natural Processing (NLP), Image Computer Vision (CV), Audio speech interaction. Multi-modal models, other hand, leverage...
Brain-computer interface (BCI) based on steady-state visual evoked potentials (SSVEP) is a popular paradigm for its simplicity and high information transfer rate (ITR). Accurate fast SSVEP decoding crucial reliable BCI performance. However, conventional methods demand longer time windows, deep learning models typically require subject-specific fine-tuning, leaving challenges in achieving optimal performance cross-subject settings. This paper proposed biofocal masking attention-based method...
Abstract Reverse water gas shift (RWGS) reaction is an attractive approach to convert CO 2 with renewable H produce CO. However, this always accompanied by undesirable methanation lowering the selectivity, particularly, at low temperature and on Ni catalysts. Herein, a strategy of deposition inert Ag surface block during RWGS for selective conversion toward was reported. Characterizations showed that in contrast formation bulk Ni−Ag alloy, both particle highly dispersed are direct contacting...
Proton exchange membrane fuel cells (PEMFCs) have achieved milestones in performance improvements and commercial launches. In the typical commercialized PEMFCs, compressed air to cathode is usually supplied from ambient air, assuming that no costly pre-purification system applied. Therefore, working PEMFCs may suffer negative effects of impurities. this regard, SO2, as most poisonous species, be fed along with at strongly adsorbed on Pt surface, leading site deactivation. To address...
Abstract An industrial robot is a complex mechatronics system, whose failure hard to diagnose based on monitoring data. Previous studies have reported various methods with deep network models improve the accuracy of fault diagnosis, which can get an accurate prediction model when amount data sample sufficient. However, obtain, leads few-shot issue and bad generalization ability model. Therefore, this paper proposes attention enhanced dilated convolutional neural (D-CNN) approach for...
Gait recognition, a task of identifying people through their walking pattern, has attracted more and researchers' attention. At present, most skeleton-based gait recognition approaches extract features from merely joint coordinates. However, the information, e.g. bone motion, is equally instructive discriminative for recognition. Thus, this paper proposes novel multi-stream part-fused graph convolutional network, MS-Gait, to fuse part-level information capture multi-order skeleton data. To...
Ambient contaminants, e.g., sulfur dioxide (SO2), lead to severe Pt deactivation and performance loss of practical proton exchange membrane (PEM) fuel cells, which need be removed immediately. We herein find that only partially adsorbed SO2 can oxidized at up 1.5 V, even for a longer period. By contrast, interestingly, dynamic potential scanning several cycles (usually > 8) gradually regenerate poisoned Pt/C. In situ infrared spectroscopy demonstrates parallel-bonded is forbidden...
Abstract Due to information asymmetry, spatial distance will affect the associated credit cooperation relationship among enterprises. This paper employed real‐world economic data construct entropy interaction complex network model, which effectively depicts risk enterprises with consideration of influence and regional development level. Based on calculation experiment Moran index, this reveals correlation contagion in each region builds theoretical model for contagion. The research findings...
Strict food regulations, high‐quality product demand, and fierce market competition urge enterprises to focus on the optimization strategy of safety traceability recall in chain. To better meet customers' quality expectations reduce loss problematic products that fail be recalled time, paper establishes a two‐level supply chain consisting single manufacturer supplier, uses Stackelberg game explore interaction between chain, finds best investment strategy. And we consider case former...
An O2O supply chain consisting of a manufacturer with an online direct channel and retailer who resells through brick-and-mortar store is considered. Three power structures (vertical Nash, Stackelberg, Stackelberg) three pricing sequences (simultaneous pricing, early, early) are Counter-intuitively, under the Stackelberg structure, has first-mover advantage retailer-pricing-early achieves Pareto optimality. In other cases, have late-mover advantage. Under vertical Nash both parties may get...