Digital Twin Logistics

Aims & Scope

The escalating demand for operational efficiency and cost optimization in logistics systems has elevated the necessity for intelligent system upgrades to unprecedented levels. Contemporary logistics ecosystems are witnessing accelerated integration of automated machinery and robotics across core operational nodes including warehousing, sorting, packaging, and transportation. This convergence of human expertise, mechanical automation, and robotic intelligence has given rise to next-generation intelligent logistics systems that prioritize collective system performance over individual component capabilities.

 

Accurate prediction of systemic performance metrics has emerged as a critical prerequisite for optimal system design and operational control. Conventional analytical methodologies demonstrate inherent limitations in addressing the dynamic complexity of these cyber-physical systems. Digital Twin (DT) technology presents a paradigm-shifting solution, enabling bidirectional synchronization between physical logistics infrastructure and their virtual counterparts through real-time data integration and simulation modeling.

 

As an innovative technological framework, digital twins facilitate unprecedented fidelity in virtual system representation, proving particularly valuable throughout the lifecycle of logistics systems - from initial design prototyping and system debugging to continuous operational optimization. Recent years have witnessed exponential growth in Logistics Digital Twin (LDT) research and industrial applications, establishing this domain as a focal point in global technological innovation.

 

This dedicated session aims to present cutting-edge advancements in LDT theory and practice, featuring interdisciplinary research that bridges academic innovation with industrial implementation. Contributions will explore novel methodologies for digital twin development, implementation challenges, and transformative applications in modern logistics ecosystems.

 

The session will focus on the following aspects:

▪ Methodologies of combing logistics system and digital twin.

▪ AI for LDT.

▪ Innovative researches of RDT

▪Advanced application researches of LDT

▪ Performance prediction of logistics system via LDT


------


Your valuable insights are welcome! 

We cordially invite interested researchers to contact us for details and presentation applications at: secretariat@idea-global.net

Session Chairs

Presentations

  • Professor
    University of Science and Technology Beijing
    Title: Digital twin driven four-way pallet shuttle storage and retrieval system optimization
    Abstract "The Four-way pallet Shuttle Storage and Retrieval System (FSS/RS) is a relatively new type of logistics equipment used in warehouses or workshops. Compared with the traditional Automated Storage and Retrieval System (AS/RS), the FSS/RS boasts advantages in flexibility and has achieved rapid development. In most cases, the FSS/RS is customized to meet different performance indicators, such as storage capacity and storage/retrieval efficiency. For this reason, optimizing the layout of the FSS/RS and the coordination of shuttles is crucial before the system is put into use. A digital twin of the FSS/RS is developed using Siemens Plant Simulation, through which different layout scenarios and shuttle coordination algorithms can be verified. According to practical applications, the FSS/RS digital twin can quickly enhance the capabilities of designers and improve the overall performance of the FSS/RS system."
  • Assistant Professor
    University of Birmingham
    Title: A Digital Twin Framework for Real-Time Railway Bridge Strike Prevention, Detection, and Structural Health Assessment
    Abstract Railway bridge strikes pose a severe threat to transport infrastructure, causing significant financial losses and operational delays. To overcome the limitations of time-consuming manual inspections, this study proposes a novel Digital Twin framework for the real-time prevention, detection, and structural health assessment of railway bridges. The architecture integrates physical edge-sensing, utilizing LiDAR and camera-based vision systems for pre-impact vehicle profiling, alongside wireless vibration sensors, with a cloud-based digital simulation layer. By leveraging a high-fidelity Finite Element (FE) model and a pre-computed database, the system rapidly maps physical collision data to evaluate structural damage and displacement. The platform translates raw sensor data into actionable operational decisions (e.g., normal, speed restriction, or immediate closure) within seconds. Successfully validated and deployed on the UK rail network, this framework provides a highly replicable, commercially viable paradigm for smart transport infrastructure and civil engineering management.
  • Associate Professor
    Southwest Jiaotong University
    Title: Multimodal Decision-Making for Intelligent Warehouse Robotics via Large Language Models and Reinforcement Learning
    Abstract Autonomous mobile robots are increasingly used in smart logistics, flexible manufacturing, and robotized material handling. This talk focuses on turn-aware multi-agent path finding for AMR coordination, addressing limitations of classical approaches that neglect orientation and in-place rotations. A scalable iterative planning framework is proposed to explicitly model turn actions, rotation delays, and local robot interactions, improving both planning realism and computational efficiency in large-scale shared environments. Experimental results and a flexible manufacturing case study show that incorporating turning constraints enhances the consistency between planned and actual execution costs and reduces operational inefficiencies compared to turn-agnostic methods. In addition, ongoing work on search-based anytime planning under turn constraints is introduced, aiming to improve high-level search control and solution refinement. Overall, the study emphasizes the importance of realistic motion modeling for efficient large-scale multi-robot coordination.
  • Assistant Professor
    Tongji University
    Title: Turn-Aware Multi-Agent Path Finding for Scalable AMR Coordination
    Abstract his talk addresses turn-aware multi-agent path finding for autonomous mobile robot coordination in smart logistics and flexible manufacturing systems. Unlike conventional methods that neglect robot orientations and turning operations, the presented work explicitly models turn actions, rotation delays, and local robot interactions to achieve more realistic and executable plans. An iterative planning framework is introduced to improve scalability and computational efficiency in large-scale multi-robot environments. Benchmark evaluations and a flexible manufacturing case study demonstrate that incorporating turn constraints enhances the consistency between planned and actual execution costs while reducing overall operational costs. The talk also outlines ongoing research on search-based anytime planning under turn constraints, focusing on improving search efficiency and solution refinement. Overall, the work emphasizes the importance of realistic motion modeling for efficient and reliable large-scale AMR coordination.
  • Professor
    Shandong Jianzhu University
    Title: Operational Optimization and Management Framework for D2C Distribution Centers Based on Digital Models
    Abstract Unlike B2C e-commerce distribution centers (DCs) that handle large-volume and full-category orders, Direct-to-Consumer (D2C) DCs focus on vertical categories with concentrated orders and limited SKUs, primarily relying on person-to-goods picking. Using a beauty and daily necessities DC as a case study, this research developed a digital model to create a virtual-to-real mapping of layouts, workflows, and order data. The model precisely diagnosed several operational bottlenecks: redundant picking paths, congestion-induced waiting times, poor cross-stage synchronization, and imbalanced staffing. To address these issues, the effectiveness of two optimization strategies was validated via the model:(1) The joint Rank Order Clustering (ROC) algorithm was adopted for order batching and slotting to reduce traveling and waiting times; (2) The Heijunka-based Variable Neighborhood Search (VNS) algorithm was applied to balance workloads across batches. During peak promotion periods, the overall order processing time was reduced by 30%–40%. Furthermore, the model enabled dynamic personnel configuration to guarantee order fulfillment within required time windows. This "Digital Diagnosis–Simulation Verification–Tactical Deduction" framework provides a feasible solution for cost savings and efficiency improvement in D2C logistics operations.
  • Associate Professor
    Jinan University
    Title: A4PS: Agentic AI-Assisted Advanced Planning and Scheduling with Large Language Models for Smart Manufacturing
    Abstract Advanced Planning and Scheduling (APS) for manufacturing systems is becoming more complex against the backdrop of intelligent transformation and increasing demand for customization. This paper proposes an Agentic AI-Assisted APS (A4PS) framework that leverages large language models and multi-agent systems to support APS modification and update processes under dynamic real-world conditions. A structured multi-agent workflow is designed to align with standard APS procedures, enabling coordinated cross-domain collaboration. A multi-step knowledge augmentation strategy is introduced to incorporate specialized APS expertise into LLM-based agents. In addition, retrieval-augmented generation and chain-of-thought reasoning are employed to enhance knowledge utilization and interaction quality. Experimental results on a constructed APS dataset show that A4PS significantly outperforms baseline LLMs in modelling success rate, solution accuracy, algorithm code correctness, and executability. A case study further demonstrates its ability to emulate expert reasoning and support APS tasks through natural language interaction, highlighting its potential for AI-driven smart manufacturing applications.
  • Postdoctoral Researcher
    Beijing Research Institute of Automation for Machinery Industry Co., Ltd.
    Title: Mirror the Mind of Crew: Digital Twin-based commissioning of scheduling and control algorithms for logistics equipment
    Abstract Addressing industry pain points such as severely compressed commissioning cycles, the lack of on-site readiness, and immense delivery pressure in logistics equipment engineering and R&D , this report proposes a Digital Twin-based virtual commissioning method for logistics equipment. By constructing high-fidelity physical models and utilizing electrical signal-driven operational logic , this approach successfully overcomes the heavy reliance of scheduling and control algorithms on physical prototypes and site environments. It enables the synchronization of software development with prototype manufacturing, facilitating an engineering model of "remote testing and local deployment". Implementation results demonstrate that this technology significantly shortens R&D and testing cycles. While enhancing product stability and iteration speed, it substantially reduces on-site labor requirements and commissioning costs, ultimately bolstering the enterprise's market competitiveness.