Digital-twin-based full-link supply chain data management method and system

Abstract

The application provides a full-link supply chain data management method and system based on digital twinning, which comprises the following steps: collecting real-time data from multiple supply chain data sources, inputting the data into a digital twinning model, and storing the data in a distributed database; performing local data preprocessing on the digital twinning model based on an edge computing node; designing a comprehensive objective function for measuring scheduling effect, dynamically adjusting the weight of each optimization objective according to real-time environmental changes and task requirements, and ensuring that each task is reasonably allocated resources according to its priority; designing a three-dimensional visualization system to display the optimized tasks and resource states in a three-dimensional space, and displaying the state of each link of the supply chain, the optimization effect and potential problems in real time through a three-dimensional visualization interface. The application adopts advanced technologies such as digital twinning technology, edge computing, adaptive optimization algorithm and cloud-edge collaboration, and solves many pain points in traditional medical supply chain management.

Description

Technical Field

[0001] This invention belongs to the field of digital twins, and particularly relates to a method and system for full-chain supply chain data management based on digital twins. Background Technology

[0002] In modern healthcare supply chain management, with the rapid growth of healthcare demand and the increasing complexity of supply chain management, ensuring the timely, accurate, and safe supply of pharmaceuticals, medical devices, and other medical supplies has become a core issue that urgently needs to be addressed. Traditional healthcare supply chains typically involve multiple stages, including production, warehousing, transportation, distribution, and inventory management, requiring rapid and efficient information and logistics support between these stages. However, existing healthcare supply chain management technologies often face several major challenges that limit their efficiency and flexibility.

[0003] First, existing supply chain management systems suffer from information silos in data processing and scheduling optimization. Many medical institutions and suppliers use independent systems, such as ERP (Enterprise Resource Planning) and SCM (Supply Chain Management), and these systems do not achieve sufficient information sharing. Information from each link exists in isolation in different systems, leading to poor data exchange, delayed decision-making, and even situations of excess or shortage of inventory at critical moments due to a lack of accurate data support. This information silo problem seriously affects the overall efficiency of the supply chain, especially in the medical industry, where timely delivery of supplies and accurate inventory management are directly related to patient treatment and safety.

[0004] Secondly, with the development of internet and IoT technologies, massive amounts of real-time data in the medical supply chain (such as inventory status, transportation routes, environmental conditions, and order information) have gradually become a crucial component of management. Despite this, traditional centralized computing architectures (such as data centers and servers) have not fully utilized edge computing technology to enhance real-time data processing capabilities. In traditional architectures, data needs to be processed and stored through central servers, often resulting in latency, especially when dealing with large amounts of real-time dynamic data, making efficient and immediate decision-making difficult. This is particularly pronounced when managing time-sensitive and high-risk materials such as pharmaceuticals and medical devices. Due to the slow response speed of traditional systems, the supply chain often cannot make timely and effective adjustments in emergencies such as demand fluctuations, transportation delays, or insufficient inventory, thus affecting the stability and emergency response capabilities of the entire supply chain.

[0005] Furthermore, as healthcare supply chain management becomes increasingly complex and diversified, achieving coordinated optimization across all stages has long been a challenge for the industry. Traditional supply chain optimization methods are mostly based on static optimization models, lacking the ability to be flexible and dynamically adjusted. For example, when a sudden event occurs in a link of the supply chain (such as insufficient storage, transportation delays, or demand fluctuations), traditional optimization methods cannot quickly respond to these dynamic changes and lack the ability to intelligently and adaptively adjust. Especially for materials such as pharmaceuticals and vaccines that have strict requirements for temperature, humidity, and transportation timeliness, traditional supply chain management systems often cannot achieve real-time monitoring and optimization, leading to the risk of inventory backlog or expiration losses.

[0006] Furthermore, traditional visualization interfaces are mostly limited to two-dimensional or static displays, making it difficult to comprehensively and intuitively reflect the multi-dimensional and dynamic changes in the supply chain. Existing visualization interfaces often only display data from a single stage, making it difficult for managers to obtain overall information on all stages in a short period of time. This inevitably leads to information asymmetry and decision-making errors when making scheduling decisions.

[0007] Therefore, although existing technologies have made some progress in certain aspects, they still cannot effectively solve a series of key issues in medical supply chain management, such as real-time performance, flexibility, collaboration, and visualization. There is an urgent need for an innovative solution that integrates advanced technologies and can comprehensively improve supply chain efficiency and response speed. Summary of the Invention

[0008] The purpose of this invention is to propose a method and system for end-to-end supply chain data management based on digital twins. It adopts cutting-edge technologies such as digital twins, edge computing, adaptive optimization algorithms, and cloud-edge collaboration to solve many pain points in traditional medical supply chain management.

[0009] To achieve the above objectives, a first aspect of the present invention provides a method for end-to-end supply chain data management based on digital twins, the method comprising the following steps:

[0010] Real-time data is collected from multiple supply chain data sources, and a digital twin model is built based on the collected data sources. Each data source transmits data to the cloud through edge computing nodes, inputs it into the digital twin model, and stores it in a distributed database.

[0011] The digital twin model undergoes local data preprocessing based on edge computing nodes, and the preprocessed data is then encrypted using symmetric encryption. After encryption, the edge computing nodes transmit the encrypted data to the cloud.

[0012] Based on real-time data and the calculation results of edge computing nodes, a comprehensive objective function is designed. This comprehensive objective function is used to measure the scheduling effect and dynamically adjust the weights of each optimization objective according to real-time environmental changes and task requirements, and ensure that each task is allocated resources reasonably according to its priority.

[0013] The design of the 3D visualization system will display the status of optimized tasks and resources in 3D space. Through the 3D visualization interface, the status, optimization effect and potential problems of each link in the supply chain will be displayed in real time. The three coordinate axes of the 3D visualization system represent task type, resource consumption and task completion timeliness, respectively.

[0014] Furthermore, the real-time data includes order information, inventory status, transportation trajectory, and environmental monitoring;

[0015] The construction of the digital twin model is based on the data source of the supply chain, which forms the state vector of each link, and the mapping function is used to construct the state vector of each link.

[0016] Furthermore, each data source transmits data to the cloud via an edge computing node, inputs it into the digital twin model, and stores it in a distributed database, including:

[0017] Each data source transmits data to the cloud through edge computing nodes. When there is a delay in the transmission of the data source, the digital twin model is updated in real time using the correction function of the data source for the state of the digital twin model, so as to ensure that the state of the digital twin model is consistent with the operational state of the real world at every moment.

[0018] If data loss, abnormal fluctuations, or noise interference occur during the actual data acquisition process, then execute:

[0019] The system triggers an early warning mechanism and automatically adjusts the state of the digital twin model to recalibrate it.

[0020] Furthermore, the local data preprocessing includes:

[0021] Data noise removal is performed based on the data source of the digital twin model; the data noise removal process is adjusted based on the deviation between the outlier values ​​of the current data and the distribution of historical data.

[0022] Based on the data source of the digital twin model, in the case of missing data, a time-series interpolation method is used to fill the missing data points to maintain the continuity and consistency of the data.

[0023] After collecting and preprocessing the data sources, the edge computing nodes are responsible for weighted aggregation of data from different sources to obtain a single dataset. In this single dataset, when the accuracy of a data source is high, its weight value will be increased accordingly, thereby ensuring that the fused data is more accurate.

[0024] Furthermore, the target loss is determined based on the comprehensive objective function;

[0025] The process of dynamically adjusting the weights of each optimization objective based on real-time environmental changes and task requirements includes:

[0026] The weighted environment evaluation function for each optimization objective evaluates the priority of different objectives based on real-time feedback and adjusts the weight of each optimization objective according to the priority of different objectives; wherein, the weighted environment evaluation function is the environmental change value evaluated based on real-time data;

[0027] In addition, after determining the target loss, it is also necessary to calculate the delay penalty term. By introducing the delay penalty term, the scheduling order of tasks can be optimized, reducing resource waste or performance degradation caused by scheduling delay.

[0028] Furthermore, the delay penalty term penalizes tasks with longer delays, forcing the system to prioritize urgent or delayed tasks and ensuring optimal resource allocation.

[0029] Furthermore, ensuring that resources are allocated reasonably according to the priority of each task includes:

[0030] During task execution, feedback data on task execution is collected in real time, and dynamic adjustments are made based on this feedback. At the same time, the comprehensive objective function and resource allocation are continuously optimized through the error function to achieve continuous optimization of the scheduling objective. The error function is used to measure the gap between the current scheduling result and the expected target. During the optimization process, the error is minimized by continuously adjusting the weight of each optimization objective, the priority of different objectives, and resource allocation, so that each scheduling is closer to the expected target, achieving the effect of self-optimization.

[0031] Furthermore, the coordinate system of the three-dimensional space is (X,Y,Z), where the X-axis represents the task type, the Y-axis represents resource consumption, and the Z-axis represents the timeliness of task completion.

[0032] The status of a task is represented by color coding: tasks with longer delays are shown in red, urgent tasks in yellow, and normal tasks in green.

[0033] Simultaneously, a dynamic priority model is designed based on the coordinate system of three-dimensional space, and the priority is adjusted by evaluating the spatial coordinates and resource usage of the task in real time.

[0034] The dynamic priority model is determined based on the basic priority of task i, the coordinates of task i in three-dimensional space, and the coordinates of task j in three-dimensional space.

[0035] The dynamic priority model can dynamically adjust task priority based on the task's location in three-dimensional space. Urgent tasks have higher priority, and tasks with the highest priority are those that are closest to other tasks in space.

[0036] Furthermore, the decision support provided by the 3D visualization system is based on optimized task scheduling data and combined with real-time feedback to make dynamic decision adjustments. It can provide suggested adjustment schemes based on the current task status and optimization goals. Moreover, through the synergy between the 3D visualization system and decision support, scheduling optimization and resource allocation decisions can be made in real time.

[0037] A second aspect of the invention provides a digital twin-based end-to-end supply chain data management system, the system comprising:

[0038] The supply chain data source acquisition unit is used to collect data from multiple supply chain data sources in real time and build a digital twin model based on the collected data sources. Each data source transmits data to the cloud through edge computing nodes, inputs it into the digital twin model, and stores it in a distributed database.

[0039] The supply chain data source analysis unit is used to perform local data preprocessing on the digital twin model based on edge computing nodes, and to perform symmetric encryption on the preprocessed data. After encryption, the edge computing nodes transmit the encrypted data to the cloud.

[0040] The resource allocation unit is used to design a comprehensive objective function based on real-time data and the calculation results of edge computing nodes. The comprehensive objective function is used to measure the scheduling effect and dynamically adjust the weight of each optimization objective according to real-time environmental changes and task requirements, and ensure that each task is reasonably allocated resources according to its priority.

[0041] The visualization display unit is used to design a three-dimensional visualization system to display the optimization tasks and resource status in three-dimensional space. Through the three-dimensional visualization interface, the status, optimization effect and potential problems of each link of the supply chain are displayed in real time. The three coordinate axes of the three-dimensional visualization system represent the task type, resource consumption and task completion timeliness, respectively.

[0042] The beneficial technical effects of the present invention are at least as follows:

[0043] Deep Integration of Digital Twins and Edge Computing: This invention is the first to organically combine digital twin technology with edge computing, constructing a real-time, end-to-end synchronized supply chain management platform. Through digital twin models, each link in the medical supply chain is modeled and monitored in real time, ensuring that all states of the supply chain (such as inventory, transportation routes, temperature, and humidity) are always dynamically reflected. Edge computing nodes are responsible for real-time analysis and processing of local data, enabling them to make optimization decisions locally in real time. This cloud-edge collaborative architecture effectively overcomes the latency problem of traditional centralized systems, greatly improving the real-time performance of data processing and decision response, ensuring rapid response to emergencies.

[0044] Application of Adaptive Optimization Algorithms: To address the real-time scheduling challenges in the dynamic healthcare supply chain, this invention introduces adaptive optimization algorithms, particularly those based on reinforcement learning and genetic algorithms, running concurrently on edge computing nodes and cloud platforms. Through continuous data learning and feedback, the system can dynamically adjust decisions regarding inventory, transportation routes, and supply plans based on actual conditions, thereby achieving intelligent optimization and adapting to changes in different scenarios. For example, the system can quickly adjust based on real-time demand fluctuations, transportation delays, or inventory status, ensuring efficient supply chain collaboration.

[0045] Intelligent Scheduling and Visualization Interface for the Entire Supply Chain: To improve the visibility and accuracy of management decisions, this invention develops a 3D visualization interface. This interface not only intuitively presents real-time data from each link in the supply chain but also displays key supply chain indicators (such as inventory levels, transportation progress, and environmental conditions) in dynamic, interactive charts. This 3D visualization helps managers grasp the overall dynamics of the supply chain in real time, enabling more comprehensive and scientific decision-making. Furthermore, based on digital twin technology, managers can simulate the operation of each link in the supply chain through a virtual interface, predict potential problems in advance, and make timely adjustments before problems occur. Attached Figure Description

[0046] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0047] Figure 1 This is a flowchart of the end-to-end supply chain data management method based on digital twins according to the present invention. Detailed Implementation

[0048] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0049] like Figure 1 As shown in the embodiment of the present invention, the end-to-end supply chain data management method based on digital twins includes:

[0050] Real-time data is collected from multiple supply chain data sources, and a digital twin model is built based on the collected data sources. Each data source transmits data to the cloud through edge computing nodes, inputs it into the digital twin model, and stores it in a distributed database.

[0051] The digital twin model undergoes local data preprocessing based on edge computing nodes, and the preprocessed data is then encrypted using symmetric encryption. After encryption, the edge computing nodes transmit the encrypted data to the cloud.

[0052] For encrypted data in the cloud, a comprehensive objective function is designed. This comprehensive objective function is used to measure the scheduling effect and dynamically adjust the weight of each optimization objective according to real-time environmental changes and task requirements, and ensure that each task is allocated resources reasonably according to its priority.

[0053] The design of the 3D visualization system will display the status of optimized tasks and resources in 3D space. Through the 3D visualization interface, the status, optimization effect and potential problems of each link in the supply chain will be displayed in real time. The three coordinate axes of the 3D visualization system represent task type, resource consumption and task completion timeliness, respectively.

[0054] Specifically, the following provides a full explanation of this embodiment:

[0055] The process involves real-time data collection from multiple supply chain data sources, constructing a digital twin model based on the collected data, wherein each data source transmits data to the cloud via edge computing nodes, inputs it into the digital twin model, and stores it in a distributed database, including:

[0056] Specifically, the core objective of this step is to establish an accurate digital twin model to ensure real-time data synchronization across all stages of the healthcare supply chain, providing accurate foundational data for subsequent optimization. The digital twin model will integrate data from various systems, such as ERP, SCM, and IoT devices, to demonstrate the actual status of the supply chain, including information on inventory, orders, and shipping progress.

[0057] First, real-time data is collected from multiple data sources. Each link in the supply chain may involve data sources from different systems, including order information (ERP system), inventory status (WMS system), transportation tracking, and environmental monitoring (IoT devices, etc.). Each data source... At this moment, t will provide key data for this stage. For example:

[0058] This indicates order status data.

[0059] Represents inventory data.

[0060] This represents data on transportation and environmental conditions.

[0061] This data is preprocessed at edge computing nodes before being transmitted to the cloud processing system. To ensure the real-time nature and reliability of the data, the edge computing nodes perform preliminary data cleaning and formatting before uniformly inputting it into the digital twin framework.

[0062] Furthermore, once the data is acquired through preprocessing and acquisition mechanisms, it is integrated into a unified digital twin model. The goal of this model is to accurately reflect the real-time status of each stage. The model is built upon multiple dimensions of the supply chain, including inventory levels, order status, and transportation progress, forming a state vector for each stage. The state vector for each stage i... The status of an inventory can be determined by the outputs of different data sources. For example, inventory status may be affected by order processing and transportation status, which in turn is affected by environmental conditions, inventory levels, etc. Therefore, the status of each stage in a digital twin model is a dynamically changing function, which can be expressed as:

[0063]

[0064] Here, f is a mapping function representing the influence of different data sources on the state of each link in the supply chain. The state of each link not only reflects the current data situation, but is also adjusted and optimized based on historical data and a preset state change model.

[0065] Furthermore, to ensure that the digital twin model reflects the latest supply chain status, this invention requires the design of an efficient real-time data transmission mechanism. Each data source transmits data to the cloud via edge computing nodes, and the data experiences a certain delay during transmission. Assuming the transmission delay is δ(t), after transmission to the cloud, the data will be synchronized and updated in time according to the delay. This mechanism ensures that the model state is consistent with the operational state of the real world at every time t. Specifically, the data source transmits the collected data to the cloud via edge computing nodes and updates the model in real time:

[0066]

[0067] in, It is an updated digital twin model. Through data source A correction function for the model state.

[0068] Furthermore, during actual data acquisition, data loss, abnormal fluctuations, or noise interference may occur. To ensure the accuracy of the digital twin model, this invention designs a real-time anomaly detection and state calibration mechanism. During data acquisition, the state data at each stage is checked for outliers, such as significant deviations in inventory quantity from expected values ​​or extreme values ​​in transportation status. If an anomaly is detected, the system triggers an early warning mechanism and automatically adjusts the state of the digital twin model to recalibrate it. Anomaly detection can be performed using a preset threshold function. To proceed:

[0069]

[0070] Once an anomaly is detected, the system will adjust the model state according to a pre-defined calibration algorithm to ensure the accuracy and reliability of the digital twin model.

[0071] Furthermore, and finally, all real-time updated data and the state of the digital twin model require efficient storage and access. In this step, a distributed storage architecture was designed to store historical data and current status of the supply chain. This architecture ensures that in subsequent steps, managers can query and review various supply chain metrics at any time and perform in-depth analysis. The storage architecture needs to support large-scale parallel data processing to cope with the complexity and volume of data in the healthcare supply chain.

[0072] The digital twin model and its associated data for each time step t will be stored as:

[0073]

[0074] This storage mechanism ensures the persistence and traceability of data.

[0075] The process of performing local data preprocessing on the digital twin model based on edge computing nodes, followed by symmetric encryption on the preprocessed data, and then transmitting the encrypted data to the cloud via the edge computing nodes, includes:

[0076] In this step, the present invention preprocesses the real-time acquired data through edge computing nodes to reduce data transmission latency and improve data processing efficiency. Edge computing can effectively filter data noise, format it, and impute missing values, thereby providing cleaner data for subsequent analysis.

[0077] Healthcare supply chain data is often subject to various interferences, such as sensor errors and environmental changes, which can lead to noise in the data. To remove this noise, this invention designs an adaptive noise removal mechanism that dynamically adjusts the denoising parameters based on the distribution of historical data and the anomalies of current data. In this step, the input dataset is the raw data from a digital twin model. The result after noise removal is The noise removal process is adjusted based on the deviation between outliers in the current data and the distribution of historical data:

[0078]

[0079] in, This represents the noise term, calculated using a dynamic algorithm based on the fluctuation range of historical data and the changes in the data at the current moment. For example, if the data value at the current moment deviates significantly from the historical average, it is considered that the data value is affected by significant noise, and adjustments are made accordingly. The calculation of the noise term is based on adaptive adjustment of the error range, ensuring more accurate noise removal.

[0080] Furthermore, data formatting ensures that data from different sources is uniformly represented under the same data structure, facilitating subsequent calculations and transmission. In the event of missing data, this invention employs a temporal interpolation method to fill in missing data points, maintaining data continuity and consistency. The input dataset for this step comes from the results processed in the previous step. The output is a formatted and interpolated dataset. Specifically, if data is missing at a certain moment, it is filled in using linear interpolation:

[0081]

[0082] Here, α(t) is the time-weighted coefficient, which is dynamically adjusted based on the length of the time interval before and after the missing data point. A smaller time interval means a more accurate interpolation result, ensuring that the filled data better reflects the actual situation.

[0083] Furthermore, in this step, the edge computing nodes are responsible for aggregating data from different sources. The outputs from multiple sensors or data sources may be temporally skewed or differ in measurement scale, thus requiring weighted processing. Assuming multiple data sources... and The weights are w i(t) and w j (t), edge computing nodes use a weighted average method to merge these data sources into a single dataset. :

[0084]

[0085] Among them, w i (t) and w j (t) is a dynamically calculated weighting coefficient, adjusted based on the quality of the data source (such as signal-to-noise ratio, real-time performance, etc.). When the accuracy of a data source is high, its weight value will be increased accordingly, thereby ensuring that the fused data is more accurate.

[0086] Furthermore, healthcare supply chain data is highly sensitive, therefore data security must be ensured. Data processed at edge computing nodes must be encrypted to prevent leakage during transmission. This invention designs a scheme based on symmetric encryption for the aggregated data. Encryption processing:

[0087]

[0088] in, Let ε be the encryption key at time t, and ε be the encryption function, ensuring that the transmitted data is stored encrypted. To enhance security, this invention employs a periodically updated key scheme to ensure long-term data privacy and security.

[0089] Furthermore, after completing data preprocessing, encryption, and aggregation, the edge computing nodes transmit the data to the cloud for further processing. During data transmission, the edge nodes compress the data to reduce bandwidth consumption and improve transmission efficiency. To ensure data integrity and accuracy, each data packet is appended with a timestamp to ensure that the data can be processed and stored in the cloud in chronological order.

[0090] Ultimately, the encrypted and compressed data is uploaded to the cloud, ensuring that the digital twin model in the healthcare supply chain system can obtain reliable and secure data in real time, providing high-quality input for subsequent scheduling and optimization.

[0091] Based on real-time data and the calculation results of edge computing nodes, a comprehensive objective function is designed. This comprehensive objective function is used to measure the scheduling effect and dynamically adjusts the weights of each optimization objective according to real-time environmental changes and task requirements, ensuring that each task is allocated resources reasonably according to its priority. This includes:

[0092] The goal of this step is to optimize the operational efficiency of the entire supply chain through intelligent scheduling based on real-time data and edge computing results, ensuring the rational allocation of resources and the smooth execution of tasks, and avoiding problems such as excess or shortage of inventory and transportation delays.

[0093] Building upon the previous step, and considering the multi-objective optimization needs in the medical supply chain (such as cost, timeliness, and resource utilization efficiency), this invention designs a comprehensive objective function. This is used to measure the effectiveness of scheduling. The objective function will include multiple dimensions such as resource utilization, cost control, and latency optimization. Based on this, this invention introduces dynamic weights to reflect the changing importance of different objectives at different times. The objective function expression is:

[0094]

[0095] Among them, w i (t) is the dynamic weight coefficient of the i-th target, which is adjusted according to real-time data changes; f i (x(t)) is the i-th objective function, representing the efficiency of task execution (such as time, resource utilization, etc.); C(t) is the total cost at the current moment, including storage, transportation, and other related costs; λ(t) is the cost adjustment factor; D(t) is the delay loss of task execution, and α(t) is the adjustment coefficient of the delay loss. This objective function provides a mathematical framework for multi-objective scheduling optimization, and the optimization objective can be adjusted in real time according to task type, demand priority, etc.

[0096] Furthermore, since resource demands in the healthcare supply chain frequently change, a flexible mechanism is needed to dynamically adjust the weights of different objectives. This invention designs an adaptive weight adjustment mechanism based on encrypted historical data from the previous stage and real-time task data.

[0097] The weight adjustment function can update the weights of each objective by evaluating the current environment. Assume this invention defines an environmental evaluation function. Based on real-time feedback, the priority of different objectives is assessed. The weight of each objective is then dynamically adjusted using this function.

[0098] The weight update formula is as follows:

[0099]

[0100] in, β is the initial weight of target i; i It is the sensitivity coefficient of the target, representing the importance of different targets to scheduling optimization; It is an environmental change value assessed based on real-time data. This mechanism allows the system to automatically increase the weight of certain objectives (such as timeliness) and optimize system response speed when demand fluctuates or urgent tasks occur.

[0101] Furthermore, the core of task scheduling lies in the rational allocation of system resources to ensure that high-priority tasks can be completed in a timely manner. To achieve this goal, this invention designs a priority-based dynamic scheduling method. Task priority... The resource allocation is calculated by combining task urgency, remaining resources, and expected execution time. The formula is as follows:

[0102]

[0103] in, It is the priority of task j at time t; This represents the total amount of resources currently available to the system. This formula ensures a direct link between task priority and resource allocation, allowing high-priority tasks to be processed first when resources are scarce. Furthermore, this mechanism incorporates real-time system feedback, further enhancing the flexibility and efficiency of resource scheduling.

[0104] Furthermore, in the medical supply chain, delays can lead to disruptions in the supply of materials, impacting the efficiency of the entire system. To mitigate the negative impact of delays on overall scheduling, this invention designs a delay penalty optimization term. This term optimizes the scheduling order of tasks by introducing a delay penalty term, reducing resource waste or efficiency degradation caused by scheduling delays. The delay penalty term is defined as follows:

[0105]

[0106] Where, δ j (t) is the delay penalty coefficient of task j at time t, representing the impact of task delay on the overall scheduling; Δt j (t) is the actual delay time of task j. This loss term penalizes tasks with longer delays, forcing the system to prioritize urgent or delayed tasks and ensuring optimal resource allocation.

[0107] Furthermore, during task execution, the system collects real-time feedback data on task execution and makes dynamic adjustments based on this feedback. This invention continuously optimizes the objective function and resource allocation through an error function to achieve continuous optimization of the scheduling objective. The error function measures the gap between the current scheduling result and the expected target; during optimization, the system attempts to minimize the error. The expression for the error function is:

[0108]

[0109] Among them, f i(x(t)) is the actual scheduling result of task i at time t; This is the expected scheduling result for task i. The system continuously adjusts the target weight, task priority, and resource allocation to make each scheduling closer to the expected goal, achieving a self-optimization effect.

[0110] Furthermore, after all tasks are scheduled, each node (such as warehouses, transportation vehicles, etc.) will execute the tasks according to the scheduling plan. During task execution, the system continuously tracks the execution status and feeds back the task status to the central control system. This end-to-end management mechanism ensures that each node can work efficiently and collaboratively to achieve optimal resource allocation.

[0111] Through continuous feedback and optimization, the system can dynamically adjust its scheduling strategy when faced with resource constraints, demand fluctuations, or changes in task urgency, ultimately ensuring the efficient operation of the entire healthcare supply chain.

[0112] The proposed 3D visualization system displays the optimization tasks and resource status in 3D space. Through a 3D visualization interface, it provides real-time updates on the status, optimization effects, and potential problems at each stage of the supply chain. The three coordinate axes of the 3D visualization system represent task type, resource consumption, and task completion timeliness, including:

[0113] It provides managers with intuitive decision support, displaying the status, optimization effects, and potential problems of each link in the supply chain in real time through a 3D visualization interface, ensuring that managers can make quick and accurate decisions in a complex supply chain environment.

[0114] In healthcare supply chain scheduling optimization, an intuitive 3D visualization system is a core component of decision support. It not only helps decision-makers understand the supply chain status in real time but also helps identify potential bottlenecks, predict future risks, and thus optimize resource allocation. Based on the output data from previous steps (such as task priorities, resource status, and delay losses), this invention designs a real-time updated 3D visualization system that displays the optimization tasks and resource status in 3D space.

[0115] The coordinate system in three-dimensional space is (X, Y, Z), where:

[0116] The X-axis represents the task type (such as drug delivery, medical equipment scheduling, etc.);

[0117] The Y-axis represents resource consumption (such as transportation vehicle utilization rate, warehouse utilization rate, etc.);

[0118] The Z-axis represents the timeliness of task completion (such as task delay time, estimated arrival time, etc.).

[0119] Task status is represented by color coding: tasks with longer delays are displayed in red, urgent tasks in yellow, and normal tasks in green. The system can receive real-time feedback from the scheduling module and adjust the visualization accordingly.

[0120] With this real-time updated information, decision-makers can quickly identify bottlenecks in the supply chain and make optimization decisions. Each time a scheduling strategy is adjusted, the system dynamically reflects the new scheduling results.

[0121] Furthermore, 3D visualization not only displays data but also requires assigning dynamic priorities to each task through a spatial state model. Based on a spatial coordinate system, this invention designs a dynamic priority model that adjusts priorities by real-time evaluation of the task's spatial coordinates and resource usage. Dynamic Priority P i The formula for calculating (t) is as follows:

[0122]

[0123] in, This is the basic priority of task i, usually determined by factors such as the urgency of the task and resource requirements; X i ,Y i Z i These are the coordinates of task i in three-dimensional space, representing the task type, resource consumption, and timeliness; X j ,Y j Z j Here, α represents the coordinates of task j in three-dimensional space; α is a coefficient that adjusts the influence of spatial distance, usually a positive number, controlling the impact of distance between tasks. This model can dynamically adjust task priority based on the task's position in three-dimensional space. Urgent tasks have higher priority, and their priority increases when they are spatially close to other tasks. This space-based priority model combines task type, resource consumption, and timeliness, accurately reflecting the relative priority between tasks.

[0124] Furthermore, the decision support provided by the 3D visualization system is based on optimized task scheduling data and incorporates real-time feedback for dynamic decision adjustments. Through the decision support module, the system can provide suggested adjustment plans based on the current task status and optimization objectives.

[0125] Suppose the system provides N scheduling tasks at time t, and the priority, resource consumption, and other information of each task are fed back to the decision support system. To maximize the scheduling effect, this invention introduces a feedback adjustment formula based on an optimization objective to automatically adjust the task priority and resource allocation strategy.

[0126] The feedback adjustment formula is as follows:

[0127]

[0128] Where, ω i (t) is the feedback weight of task i, which is dynamically adjusted based on the task type and current resource status; P i (t) is the current priority of task i at time t; This is the expected target priority for task i. Based on the feedback value... The system automatically optimizes the scheduling scheme. If the feedback value is positive, the system increases the priority or resource allocation of the task; otherwise, it decreases it. Through this mechanism, the system can adjust resource allocation in real time to respond to changes in tasks, thereby maintaining the overall scheduling efficiency.

[0129] Furthermore, through the synergy of 3D visualization and decision support systems, scheduling optimization and resource allocation decisions can be made in real time. Specifically, 3D visualization displays the timeliness, resource consumption, and spatial location of each task, while the decision support system adjusts task priorities and resource allocation based on real-time data to ensure optimal utilization of resources throughout the supply chain.

[0130] Building on this, the system's collaborative mechanism helps decision-makers quickly identify problem nodes in the supply chain and respond promptly. For critical tasks, the system increases priority, optimizes resource allocation, and reduces latency; for non-critical tasks, the system avoids excessive resource consumption through appropriate resource sharing and latency management.

[0131] Furthermore, to maintain continuous optimization of the scheduling system, a closed-loop feedback mechanism is employed. After each adjustment of resource or task priorities, the system performs real-time evaluation based on the new scheduling results and adjusts the optimization objective accordingly. The closed-loop optimization formula is as follows:

[0132]

[0133] in, It is the objective function value of the current scheduling scheme; γ is the feedback adjustment value; γ is the adjustment factor used to control the degree of influence of feedback adjustment on the optimization objective. Through continuous closed-loop optimization, the system can continuously adjust the scheduling strategy in a changing environment, ensuring the efficient operation of the medical supply chain.

[0134] Ultimately, the synergistic effect of 3D visualization and optimization feedback provides decision-makers with a comprehensive decision support platform. This platform can display the status of each task in real time and adjust optimization plans based on feedback, enabling the healthcare supply chain to respond rapidly to resource bottlenecks, task changes, and urgent needs, ensuring the long-term sustainable and efficient operation of the system. This system can flexibly cope with uncertainty and dynamic changes, providing the healthcare supply chain with a forward-looking and actionable decision support solution.

[0135] This invention also provides a digital twin-based end-to-end supply chain data management system, the system comprising:

[0136] The supply chain data source acquisition unit is used to collect data from multiple supply chain data sources in real time and build a digital twin model based on the collected data sources. Each data source transmits data to the cloud through edge computing nodes, inputs it into the digital twin model, and stores it in a distributed database.

[0137] The supply chain data source analysis unit is used to perform local data preprocessing on the digital twin model based on edge computing nodes, and to perform symmetric encryption on the preprocessed data. After encryption, the edge computing nodes transmit the encrypted data to the cloud.

[0138] The resource allocation unit is used to design a comprehensive objective function for encrypted data in the cloud. The comprehensive objective function is used to measure the scheduling effect and dynamically adjust the weight of each optimization objective according to real-time environmental changes and task requirements, and ensure that each task is reasonably allocated resources according to its priority.

[0139] The visualization display unit is used to design a three-dimensional visualization system to display the optimization tasks and resource status in three-dimensional space. Through the three-dimensional visualization interface, the status, optimization effect and potential problems of each link of the supply chain are displayed in real time. The three coordinate axes of the three-dimensional visualization system represent the task type, resource consumption and task completion timeliness, respectively.

[0140] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0141] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or units may be electrical, mechanical, or other forms.

[0142] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0143] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A full-chain supply chain data management method based on digital twins, characterized in that: The method includes the following steps: Real-time data is collected from multiple supply chain data sources, and a digital twin model is built based on the collected data sources. Each data source transmits data to the cloud through edge computing nodes, inputs it into the digital twin model, and stores it in a distributed database. The digital twin model undergoes local data preprocessing based on edge computing nodes, and the preprocessed data is then encrypted using symmetric encryption. After encryption, the edge computing nodes transmit the encrypted data to the cloud. The process involves collecting pre-processed data sources, with edge computing nodes responsible for weighted aggregation of data from different sources to obtain a single dataset. Within this single dataset, the weight of a data source with higher accuracy is increased to ensure greater accuracy of the fused data. Based on this single dataset, encryption is performed using an encryption key. Based on real-time data and the calculation results of edge computing nodes, a comprehensive objective function is designed. This comprehensive objective function is used to measure the scheduling effect and dynamically adjust the weights of each optimization objective according to real-time environmental changes and task requirements, and ensure that each task is allocated resources reasonably according to its priority. Based on the comprehensive objective function and priorities, a 3D visualization system is designed to display the optimization tasks and resource status in a 3D space. Through the 3D visualization interface, the status, optimization effects, and potential problems of each link in the supply chain are displayed in real time. The three coordinate axes of the 3D visualization system represent the task type, resource consumption, and task completion timeliness, respectively. The target loss is determined based on the comprehensive objective function; the comprehensive objective function includes several objective functions, total cost, and task execution delay loss; and the objective function represents the efficiency of task execution. The process of dynamically adjusting the weights of each optimization objective based on real-time environmental changes and task requirements includes: The weighted environment evaluation function for each optimization objective evaluates the priority of different objectives based on real-time feedback and adjusts the weight of each optimization objective according to the priority of different objectives; wherein, the weighted environment evaluation function is the environmental change value evaluated based on real-time data; At the same time, after determining the target loss, it is also necessary to calculate the delay penalty term. By introducing the delay penalty term, the scheduling order of tasks can be optimized, reducing resource waste or performance degradation caused by scheduling delay. The delay penalty applies to tasks with longer delays, forcing the system to prioritize urgent or delayed tasks and ensuring optimal resource allocation.

2. The end-to-end supply chain data management method based on digital twins according to claim 1, characterized in that, The real-time data includes order information, inventory status, transportation trajectory, and environmental monitoring. The construction of the digital twin model is based on the data source of the supply chain, which forms the state vector of each link, and the mapping function is used to construct the state vector of each link.

3. The end-to-end supply chain data management method based on digital twins according to claim 2, characterized in that, Each data source transmits data to the cloud via an edge computing node, inputs it into the digital twin model, and stores it in a distributed database, including: Each data source transmits data to the cloud through edge computing nodes. When there is a delay in the transmission of the data source, the digital twin model is updated in real time using the correction function of the data source for the state of the digital twin model, so as to ensure that the state of the digital twin model is consistent with the operational state of the real world at every moment. If data loss, abnormal fluctuations, or noise interference occur during the actual data acquisition process, then execute: The system triggers an early warning mechanism and automatically adjusts the state of the digital twin model to recalibrate it.

4. The end-to-end supply chain data management method based on digital twins according to claim 1, characterized in that, The local data preprocessing includes: Data noise removal is performed based on the data source of the digital twin model; the data noise removal process is adjusted based on the deviation between the outlier values ​​of the current data and the distribution of historical data. Based on the digital twin model, when data is missing, a time-series interpolation method is used to fill in the missing data points, maintaining the continuity and consistency of the data.

5. The end-to-end supply chain data management method based on digital twins according to claim 1, characterized in that, Ensuring that resources are allocated appropriately according to the priority of each task includes: During task execution, feedback data on task execution is collected in real time, and dynamic adjustments are made based on this feedback. At the same time, the comprehensive objective function and resource allocation are continuously optimized through the error function to achieve continuous optimization of the scheduling objective. The error function is used to measure the gap between the current scheduling result and the expected target. During the optimization process, the error is minimized by continuously adjusting the weight of each optimization objective, the priority of different objectives, and resource allocation, so that each scheduling is closer to the expected target, achieving the effect of self-optimization.

6. The end-to-end supply chain data management method based on digital twins according to claim 5, characterized in that, The coordinate system of the three-dimensional space is in, The axis represents the task type. The axis represents resource consumption. The axis represents the timeliness of task completion; The status of a task is represented by color coding: tasks with longer delays are shown in red, urgent tasks in yellow, and normal tasks in green. At the same time, a dynamic priority model is designed based on the coordinate system of three-dimensional space, and the priority is adjusted by evaluating the spatial coordinates and resource consumption of the task in real time. The dynamic priority model is based on tasks. Basic priorities and tasks Coordinates in three-dimensional space, and the task Determining coordinates in three-dimensional space; The dynamic priority model can dynamically adjust the priority of tasks based on their location in three-dimensional space. Urgent tasks have higher priority, and tasks with the lowest spatial distance from other tasks have the highest priority.

7. The end-to-end supply chain data management method based on digital twins according to claim 6, characterized in that, The decision support provided by the 3D visualization system is based on optimized task scheduling data and combined with real-time feedback to make dynamic decision adjustments. It can provide suggested adjustment schemes based on the current task status and optimization goals. Furthermore, through the synergy between the 3D visualization system and decision support, scheduling optimization and resource allocation decisions can be made in real time.

8. The end-to-end supply chain data management system based on digital twins as described in claim 1, characterized in that, The system includes: The supply chain data source acquisition unit is used to collect data from multiple supply chain data sources in real time and build a digital twin model based on the collected data sources. Each data source transmits data to the cloud through edge computing nodes, inputs it into the digital twin model, and stores it in a distributed database. The supply chain data source analysis unit is used to perform local data preprocessing on the digital twin model based on edge computing nodes, and to perform symmetric encryption on the preprocessed data. After encryption, the edge computing nodes transmit the encrypted data to the cloud. The resource allocation unit is used to design a comprehensive objective function based on real-time data and the calculation results of edge computing nodes. The comprehensive objective function is used to measure the scheduling effect and dynamically adjust the weight of each optimization objective according to real-time environmental changes and task requirements, and ensure that each task is reasonably allocated resources according to its priority. The visualization display unit is used to design a three-dimensional visualization system to display the optimization tasks and resource status in three-dimensional space. Through the three-dimensional visualization interface, the status, optimization effect and potential problems of each link of the supply chain are displayed in real time. The three coordinate axes of the three-dimensional visualization system represent the task type, resource consumption and task completion timeliness, respectively.

Images