Supply chain full-process intelligent monitoring and dynamic scheduling optimization method based on Internet of Things and digital twinning

Abstract

The invention discloses a supply chain full-process intelligent monitoring and dynamic scheduling optimization method based on the Internet of Things and digital twinning, and the method comprises the following steps: constructing a lightweight twinning body for a supply chain physical entity, binding the twinning body with a sensor corresponding to the entity, and storing the basic data of the entity; carrying out local data processing on each twinborn body through an edge computing node, and constructing a distributed training network by utilizing federated learning; the twinborn body monitors sensor data through a self-diagnosis module, and when data drift is detected, calibration parameters are generated and sensor output is corrected; according to the method, local anomaly detection and self-calibration of sensor data are realized through lightweight twin and edge computing node deployment: a data feature space is constructed by using a variational auto-encoder to identify drift, and spatial correlation of adjacent nodes is analyzed in combination with a graph neural network to realize data mutual verification, so that centralized monitoring dependence is broken through; the problem of data distortion caused by faults of a single sensor is solved.

Description

Technical Field

[0001] The present invention belongs to the field of supply chain technology, and in particular relates to a method for intelligent monitoring and dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins. Background Art

[0002] With the development of the Internet of Things (IoT) and digital twin technologies, supply chain management is transitioning from being driven by human experience to one driven by data intelligence. Existing technologies generally employ a centralized architecture for supply chain monitoring, deploying sensor networks to collect physical entity status data and uploading it to cloud servers for centralized processing and analysis. Dynamic scheduling, on the other hand, relies on static rule engines or single-objective optimization algorithms, incorporating historical data to pre-determine scheduling strategies for order allocation, inventory management, and route planning. While these solutions have improved supply chain efficiency to a certain extent, their technical architecture and implementation methods remain grounded in traditional IT architectures and have yet to fully unlock the synergistic potential of the IoT and digital twins.

[0003] However, from the perspective of existing technologies, in the field of intelligent monitoring, the centralized architecture leads to the system's high dependence on cloud servers. A single sensor failure or network interruption can easily cause data distortion or even monitoring failure, and the centralized transmission and processing of massive data increases latency and energy consumption. Traditional anomaly detection methods only perform threshold judgments on single sensor data, lack analysis of the spatial correlation between adjacent nodes, and are difficult to identify isolated false alarms or data missing. In particular, monitoring reliability is insufficient in complex scenarios such as ocean shipping and remote warehousing.

[0004] In the field of dynamic scheduling, static rule engines cannot respond to dynamic constraints such as traffic congestion, weather changes, and production capacity fluctuations in real time. Single-objective optimization makes it difficult to balance the interests of multiple parties in the supply chain, and lacks the ability to quantitatively manage ESG indicators such as carbon emissions. In addition, traditional scheduling systems rely on manual configuration strategies, and human-computer interaction is limited to viewing data reports. Digital twin-based visual intervention and real-time impact assessment cannot be achieved, resulting in inefficient decision-making in emergency scenarios. Therefore, a full-process intelligent monitoring and dynamic scheduling optimization method for the supply chain based on the Internet of Things and digital twins is proposed. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for intelligent monitoring and dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins, which solves the problems of existing supply chain monitoring relying on the cloud, low reliability, difficult scheduling, and balancing dynamic constraints.

[0006] To achieve the above object, the present invention provides the following technical solutions: A method for intelligent monitoring of the entire supply chain process based on the Internet of Things and digital twins, which implements distributed self-calibration monitoring based on the Internet of Things and digital twin technologies, includes the following steps: S101: Build a lightweight twin for a physical entity in the supply chain, where the twin is bound to the sensor of the corresponding entity and stores basic data of the entity; S102: Each twin performs local data processing through edge computing nodes and uses federated learning to build a distributed training network; S103: The twin monitors sensor data through the self-diagnosis module and generates calibration parameters and corrects sensor output when data drift is detected; S104: Adjacent twins autonomously monitor the network, analyze spatial correlation data based on graph neural networks, identify sensor anomalies and trigger the mutual verification mechanism; S105: The calibrated and verified data is synchronized to the master twin in the cloud to form a full-process monitoring system.

[0007] Preferably, a method for dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins includes the following steps: S201: Input the calibrated monitoring data into the dynamic scheduling engine to build a spatiotemporal network model including supply chain nodes and transportation routes; S202: Use a multi-agent game algorithm to balance multiple constraints and generate a global optimization solution for order allocation, inventory transfer, and path planning; S203: Integrate multi-objective optimization functions into the scheduling objectives to support the generation of scheduling strategies with different priorities; S204: Automatically optimize scheduling model parameters based on the deviation between the scheduling execution result and the plan; S205: Provide a human-computer interaction interface to support scheduling interventions and provide real-time feedback on impact assessments.

[0008] Preferably, the physical entities include production equipment, storage units, transport vehicles and logistics containers; the basic data of the entities include three-dimensional models, historical operation data and sensor calibration parameters.

[0009] Preferably, the self-diagnosis module uses a variational autoencoder to construct a normal data feature space, and detects data drift by calculating the deviation between real-time data and the feature space.

[0010] Preferably, the collaborative monitoring network is autonomous through short-range communication protocols such as Bluetooth Mesh and Zigbee, and uses Pearson correlation coefficient or spatiotemporal consistency to analyze the correlation of adjacent node data.

[0011] Preferably, the sensor data drift includes sensor aging deviation, vibration baseline drift or zero drift, and the calibration parameters are written into the sensor register via an industrial protocol such as Modbus / TCP.

[0012] Preferably, the spatiotemporal network model uses factories, warehouses, and distribution centers as nodes and transportation routes as edges, and the edge attributes include dynamic transportation time consumption and vehicle constraint parameters.

[0013] Preferably, the multi-agent game algorithm adopts a reinforcement learning model, and the twins of each participant output scheduling constraints such as production capacity data and cost-effectiveness curves.

[0014] Preferably, the multi-objective optimization function includes cost, timeliness, and carbon emission targets, and a multi-objective genetic algorithm such as NSGA-III is used to generate a Pareto optimal solution set.

[0015] Preferably, the human-computer interaction interface supports three-dimensional map dragging operations or natural language instructions, and displays the delivery risk and cost changes after scheduling intervention in real time.

[0016] The technical effects and advantages of the present invention's method for intelligent monitoring and dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins are as follows: 1. This invention realizes local anomaly detection and self-calibration of sensor data through the deployment of lightweight twins and edge computing nodes: it uses variational autoencoders (VAE) to construct data feature space to identify drift, and combines graph neural networks (GNN) to analyze the spatial correlation of adjacent nodes to achieve data mutual verification, breaking the dependence on centralized monitoring, solving the data distortion problem caused by single sensor failure, forming autonomous collaborative monitoring capabilities in disconnected environments, and significantly improving the reliability of supply chain monitoring.

[0017] 2. This invention models supply chain nodes and transportation routes based on the spatiotemporal graph neural network (ST-GNN), embeds dynamic parameters such as real-time traffic and vehicle constraints, and realizes multi-dimensional collaborative optimization of order distribution, inventory allocation, and path planning; introduces a multi-agent reinforcement learning algorithm to transform the twins of each participant into intelligent agents, balances multiple constraints such as production capacity, cost, and timeliness through a game mechanism, generates a global optimal scheduling plan covering the entire chain, and promotes the upgrade of the supply chain from static planning to dynamic intelligent decision-making.

[0018] 3. This invention integrates multi-objective optimization functions of cost, timeliness, and carbon emissions into scheduling objectives, uses algorithms such as NSGA-III to generate a Pareto optimal solution set, and supports flexible configuration of priority strategies to meet ESG compliance requirements; combines meta-learning technology to extract historical scheduling error patterns, and automatically optimizes model parameters to quickly adapt to emergency scenarios such as epidemic lockdowns and port strikes. It can achieve self-evolution of scheduling strategies without a large amount of labeled data, significantly improving decision-making efficiency in complex environments.

[0019] 4. This invention constructs a layered architecture of "Internet of Things layer-edge layer-network layer-application layer": the edge layer deploys lightweight twins and self-powered sensors to reduce hardware dependence, the network layer uses a federated learning alliance chain to ensure data privacy and cross-domain collaboration, and the application layer integrates carbon footprint accounting and visual interaction modules to improve management transparency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of a monitoring method for a full-process intelligent monitoring and dynamic scheduling optimization method of a supply chain based on the Internet of Things and digital twins proposed by the present invention; Figure 2 This is a flow chart of a dynamic scheduling method for intelligent monitoring and dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins proposed in the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention. It should be noted that, in this document, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or apparatus. In the absence of further restrictions, the elements defined by the sentence "include..." do not exclude the existence of other identical elements in the process, method, article or apparatus that includes the elements. Example 1 refer to Figure 1 This embodiment provides a method for intelligent monitoring of the entire supply chain process based on the Internet of Things and digital twins, which is used to implement intelligent monitoring of production equipment. The specific implementation content includes: Implementation scenario: An automotive parts manufacturer's engine block production line (five DMG MORI CMX600 machining centers) requires monitoring temperature, vibration, and tool wear sensor data.

[0022] Implementation steps: S101: Building a Lightweight Twin: Physical entity binding: Each machining center is bound to three types of sensors: temperature sensor, vibration sensor, and tool wear sensor; Twin data: 3D model, lightweight model exported in STL format using SolidWorks; Basic data: "temperature-vibration-wear" correlation curve and initial sensor calibration parameters stored for the past year; S102: Edge Computing and Federated Learning: Edge node deployment: Each machining center is equipped with a Raspberry Pi 4B edge terminal and the Flower federated learning framework installed; Distributed training: Local tasks, where each edge node trains a "temperature-vibration-wear" prediction model based on historical data (inputting temperature and vibration values ​​and outputting tool wear prediction values); In federated aggregation, 5 edge nodes upload model parameters to the central server each round. After 10 rounds of iteration, the global model MSE loss is reduced to 0.03μm (accuracy 92%). The central server uses weighted average aggregation parameters, and the formula is: , ,in, is the federated learning training round, is the local data volume of the kth edge node, is the model weight matrix of the kth edge node after the tth round of training, is the model bias vector of the kth edge node after the tth round of training, is the total data volume of all edge nodes; S103: Self-diagnosis and sensor calibration: The self-diagnosis module uses VAE to construct a normal data feature space; data drift detection and calibration: at 2:00 PM on a certain day, the vibration sensor's real-time data reconstruction error was 0.07 (exceeding the threshold), which was determined to be baseline drift (due to loose equipment anchor bolts); Generate calibration parameters (offset -0.03g) and write them into the sensor register via Modbus / TCP protocol. After correction, the data reconstruction error is reduced to 0.04; S104: Autonomous Monitoring Network and Mutual Verification: Network autonomy: 5 machining centers are autonomously connected to a Mesh network using the Zigbee 3.0 protocol (communication radius 100m, hop count ≤ 2); Spatial correlation analysis: The Pearson correlation coefficient of the vibration data of adjacent nodes (No. 1 and No. 2) was calculated (r=0.89, p<0.01), and a strong correlation was determined; Mutual verification mechanism: If the vibration data of node 1 is abnormal (e.g., jumps to 5g), the data of node 2 is called for verification (the vibration value of node 2 is 0.5g), confirming that sensor 1 is faulty and marking it (the missed detection rate of the traditional method is 15% → the present invention is 2%). S105: Data synchronization to the cloud: The calibrated data (temperature 28°C, vibration 0.5g, wear 12μm) is uploaded to the Alibaba Cloud ECS cloud master twin via MQTT; The master twin displays a real-time 3D view of the production line, marking the health status of the equipment (green: normal; yellow: maintenance required).

[0023] Implementation effect: The sensor data misjudgment rate has been reduced from 30% to 5% (due to twin binding and self-calibration); Federated learning training time is shortened from 2 hours / round to 15 minutes (edge ​​computing + parameter aggregation); Calibration response time is less than 10 seconds (traditional manual calibration requires 2 hours of downtime); Unplanned equipment downtime is reduced from 8 hours / month to 1 hour (global monitoring and timely warning).

[0024] Example 2 refer to Figure 2 This embodiment provides a supply chain full-process dynamic scheduling optimization method based on the Internet of Things and digital twins, which is used to implement warehousing-transportation dynamic scheduling optimization. The specific implementation content includes: Implementation scenario: A nationwide FMCG warehousing network (10 regional warehouses, 50 city warehouses) optimizes order distribution and transportation routes during promotional events.

[0025] Implementation steps: S201: Constructing spatiotemporal network model: Node definition: Factory nodes (Shanghai and Guangzhou, production capacity 10,000 pieces / day and 8,000 pieces / day); Warehouse nodes (5,000 pieces in stock in North China warehouse, 3,000 pieces in stock in East China warehouse); Distribution center nodes (Beijing, Shanghai, Guangzhou, daily demand 2,000 pieces, 3,000 pieces, and 1,500 pieces).

[0026] Edge attributes: Factory → Regional Warehouse (Highway, 12 hours, maximum load of 500 pieces / vehicle); Regional warehouse → City warehouse (railway, takes 24 hours, limited to 2,000 pieces / carriage); City warehouse → distribution center (city delivery, takes 4 hours, limited to 100 pieces / vehicle); S202: Multi-agent game optimization, participating agents: Factory agent (output capacity constraint: Shanghai factory can only ship a maximum of 8,000 pieces per day); Warehouse agent (output inventory constraint: 3,000 pieces can be allocated from the North China warehouse); Transport agent (output cost-efficiency curve: 0.5 yuan / item / 100 kilometers for highway, 12 hours).

[0027] Reinforcement Learning Model: State space (inventory at each node, goods in transit, order demand); Action space (order allocation ratio, transportation method selection); Reward function (R = -0.4 × cost - 0.3 × time - 0.3 × carbon emissions).

[0028] Optimization results: A global solution was generated (Shanghai factory → North China warehouse for 3,000 rail shipments, Guangzhou factory → East China warehouse for 2,500 road shipments), reducing total costs by 15% (80,000 yuan / month) and shortening average delivery time by 8 hours. S203: Generation of multi-objective scheduling strategies: Multi-objective function: cost (transportation + warehousing), timeliness (the longest order duration), and carbon emissions (0.1kgCO2 / piece / 100km for roads and 0.03kgCO2 / piece / 100km for railways).

[0029] NSGA-III algorithm: 100 populations, 50 iterations, generating 12 Pareto optimal strategies (e.g., Strategy A: cost 120,000 yuan, timeliness 48 hours, carbon emissions 2.5 tons; Strategy B: cost 140,000 yuan, timeliness 36 hours, carbon emissions 3.2 tons); S204: Automatic optimization of model parameters: Deviation correction: The transportation time of a certain batch exceeded the plan by 6 hours (due to traffic jam), so the road transportation time parameter was adjusted (12 hours → 15 hours); Model update: Retrained the reinforcement learning model, and the prediction accuracy increased from 82% to 90%; S205: Human-computer interaction intervention: Deviation correction: The transportation time of a certain batch exceeded the plan by 6 hours (due to traffic jam), so the road transportation time parameter was adjusted (12 hours → 15 hours); Model update: Retrained the reinforcement learning model, and the prediction accuracy increased from 82% to 90%.

[0030] Implementation effect: Scheduling calculation time is shortened from 2 hours to 3 minutes (supporting an average of 100,000 orders per day); The proportion of orders delivered within 48 hours increased from 60% to 85%; Monthly average customer complaints decreased by 70% (due to route delays); Improved decision-making efficiency for managers (intervention response time from 30 minutes to 2 minutes).

[0031] Example 3 This embodiment provides a method for intelligent monitoring and dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins, which is used to implement cross-border logistics container monitoring. The specific implementation content includes: Implementation scenario: Monitoring the temperature, humidity, tilt angle, and seal status of a cross-border e-commerce company's shipping containers (40-foot refrigerated containers).

[0032] Implementation steps: Twin construction: Binds a Sensirion SHT31 temperature and humidity sensor (±0.3°C / ±2%RH), a Bosch BMA423 tilt sensor (±180°), and a Laird TCK-300 electronic seal (with tamper alarm); stores historical transportation data and sensor calibration parameters (initial temperature and humidity drift 0.1°C / 1%RH); Autonomous network and mutual verification: 10 containers are autonomously networked via the Bluetooth Mesh protocol (number of hops ≤ 3), calculating the spatiotemporal consistency of temperature and humidity data from adjacent containers (time delay ≤ 2 seconds, deviation ≤ 5%). In the event of an anomaly, data from adjacent containers is called for verification (for example, if the temperature and humidity at C01 suddenly rise by 10°C, the data from C02 is verified to confirm whether it is due to a refrigeration failure).

[0033] Implementation effect: The accuracy of temperature and humidity data has been improved from 85% to 98% (self-calibration); The time to detect tilt anomalies has been reduced from 2 hours to 10 seconds (Mesh network mutual verification); The missed detection rate of seal removal inspection has been reduced from 20% to 5% (multi-sensor mutual verification).

[0034] Example 4 This embodiment provides a method for intelligent monitoring and dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins, which is used to implement carbon emission optimization for multi-objective scheduling. The specific implementation content includes: Implementation scenario: Battery transportation scheduling for a new energy vehicle company (priority: carbon emissions > timeliness > cost).

[0035] Implementation steps: Adjustment of the multi-objective function: Weight R = -0.5 × carbon emissions - 0.3 × time efficiency - 0.2 × cost. Carbon emissions calculation includes road (0.2 kg CO2 / package / 100 km), rail (0.05 kg CO2 / package / 100 km), and maritime (0.01 kg CO2 / package / 100 km). NSGA-III generation strategy: The optimal strategy (batteries shipped from Ningde to Shanghai Port by sea to Rotterdam and then distributed by rail to European warehouses) has a carbon emission of 1.2 tons (60% lower than the road solution), a delivery time of 15 days (extended by 3 days), and an 8% increase in cost.

[0036] Implementation effect: Carbon emissions meet EU carbon tariff requirements (carbon intensity ≤ 0.1kgCO2 / unit / 100km); Customer satisfaction with the “green supply chain” increased by 35%.

[0037] Example 5 This embodiment provides a method for intelligent monitoring and dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins, which is used to implement closed-loop verification of the entire vaccine cold chain process. The specific implementation content includes: Implementation scenario: Cold chain transportation of vaccines for a pharmaceutical company (requires 2-8°C throughout the entire process).

[0038] Implementation steps: Monitoring: The refrigerated truck's temperature sensor detected data drift (drift +1.5°C), with a VAE reconstruction error of 0.06, generating calibration parameters (corrected temperature 2.8°C). Scheduling end: The data input engine was calibrated and it was found that the original route (Highway A, 6 hours) might exceed the temperature due to traffic jam (forecasted to be 8 hours). The route was switched (Highway B, 7 hours) and 500 vaccines were allocated from a nearby warehouse.

[0039] Implementation effect: The risk of temperature exceeding the standard during the entire vaccine production process was reduced from 10% to 1% (avoiding a loss of 500,000 yuan for a single batch); The whole process response time is less than 15 minutes (traditional methods take 2 hours); The on-time delivery rate increased from 90% to 98% (customer complaints were eliminated).

[0040] Comparative Example 1 This comparative example provides a traditional supply chain monitoring and scheduling method, which specifically includes the following: Traditional methods: Monitoring: Manual inspection (2 hours / time), sensor data uploaded to the cloud via 4G (no self-calibration, false alarm rate 30%; average monthly downtime due to false alarms due to data drift is 2 times, resulting in losses of 100,000 yuan).

[0041] Scheduling: Allocate orders based on experience (the North China warehouse has an average monthly backlog of 2,000 pieces, with a storage cost of +30,000 yuan); static route planning (the average monthly overdue order rate is 15%, with a compensation cost of 50,000 yuan).

[0042] Compared with Examples 1-5 and Comparative Example 1, the full-process intelligent monitoring and dynamic scheduling system of the supply chain based on the Internet of Things and digital twin technology is significantly superior to traditional methods in terms of technical solutions, core indicators and application effects. The integration of technologies such as digital twins, federated learning, and graph neural networks in the supply chain field has the following advantages: Full-process closed-loop intelligence: From sensor data self-calibration to dynamic optimization of scheduling strategies, it realizes an automated closed loop of "monitoring-analysis-decision-execution", reducing the risk of manual intervention and misjudgment.

[0043] Multi-dimensional value enhancement: comprehensive optimization of indicators such as data accuracy, response speed, cost control, and green and low-carbon development to meet the needs of multiple industries such as manufacturing, fast-moving consumer goods, and pharmaceuticals, with significant advantages in high-compliance scenarios.

[0044] Flexible and scalable architecture: Supports multiple communication protocols such as Bluetooth Mesh and Zigbee, is compatible with industrial interfaces such as Modbus / TCP, and its lightweight twin design is suitable for enterprises of different sizes, with broad application prospects.

[0045] The above embodiments may be implemented in whole or in part through software, hardware, firmware or any other combination. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0046] Those skilled in the art will appreciate that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0047] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0048] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited to this. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

[0049] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A full-process intelligent monitoring method for the supply chain based on the Internet of Things and digital twins, which realizes distributed self-calibration monitoring based on the Internet of Things and digital twin technologies, characterized by: The following steps are involved: S101: Build a lightweight twin for a physical entity in the supply chain, where the twin is bound to the sensor of the corresponding entity and stores basic data of the entity; S102: Each twin performs local data processing through edge computing nodes and uses federated learning to build a distributed training network; S103: The twin monitors sensor data through the self-diagnosis module and generates calibration parameters and corrects sensor output when data drift is detected; S104: Adjacent twins autonomously monitor the network, analyze spatial correlation data based on graph neural networks, identify sensor anomalies and trigger the mutual verification mechanism; S105: The calibrated and verified data is synchronized to the cloud master twin to form a full-process monitoring system.

2. A dynamic scheduling optimization method for the entire supply chain process based on the Internet of Things and digital twins, characterized by: The steps include: S201: Input the calibrated monitoring data into the dynamic scheduling engine to build a spatiotemporal network model including supply chain nodes and transportation routes; S202: Use a multi-agent game algorithm to balance multiple constraints and generate a global optimization solution for order allocation, inventory transfer, and path planning; S203: Integrate multi-objective optimization functions into the scheduling objectives to support the generation of scheduling strategies with different priorities; S204: Automatically optimize scheduling model parameters based on the deviation between the scheduling execution result and the plan; S205: Provide a human-computer interaction interface to support scheduling interventions and provide real-time feedback on impact assessments.

3. The method for intelligent monitoring of the entire supply chain process based on the Internet of Things and digital twins as claimed in claim 1, characterized in that: The physical entities include production equipment, storage units, transport vehicles and logistics containers; the basic data of the entities include three-dimensional models, historical operation data and sensor calibration parameters.

4. The method for intelligent monitoring of the entire supply chain process based on the Internet of Things and digital twins as claimed in claim 1, characterized in that: The self-diagnosis module uses a variational autoencoder to construct a normal data feature space and detects data drift by calculating the deviation between real-time data and the feature space.

5. The method for intelligent monitoring of the entire supply chain process based on the Internet of Things and digital twins as claimed in claim 1, characterized in that: The collaborative monitoring network is autonomous through Bluetooth Mesh and Zigbee short-range communication protocols, and uses Pearson correlation coefficient or spatiotemporal consistency to analyze the correlation of adjacent node data.

6. The method for intelligent monitoring of the entire supply chain process based on the Internet of Things and digital twins as claimed in claim 1, characterized in that: The sensor data drift includes sensor aging deviation, vibration baseline drift or zero drift, and the calibration parameters are written into the sensor register via industrial protocols such as Modbus / TCP.

7. The method for dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins as described in claim 2 is characterized in that: The spatiotemporal network model uses factories, warehouses, and distribution centers as nodes and transportation routes as edges. The edge attributes include dynamic transportation time consumption and vehicle constraint parameters.

8. The method for dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins as described in claim 2 is characterized in that: The multi-agent game algorithm adopts a reinforcement learning model, and each participating twin outputs scheduling constraints such as production capacity data and cost-effectiveness curves.

9. The method for dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins as described in claim 2 is characterized in that: The multi-objective optimization function includes cost, timeliness, and carbon emission targets, and uses a multi-objective genetic algorithm such as NSGA-III to generate a Pareto optimal solution set.

10. The method for dynamic scheduling optimization of the entire supply chain process based on the Internet of Things and digital twins as claimed in claim 2, characterized in that: The human-computer interaction interface supports three-dimensional map dragging operations or natural language commands, and displays delivery risks and cost changes after scheduling intervention in real time.

Images