A service platform for precise supply chain inventory management

By combining data fusion, digital twin models, and intelligent decision engines, the problems of data silos and multi-objective optimization in supply chain inventory management are solved. It achieves a Pareto optimal balance between a globally consistent data foundation and multi-objective optimization, possesses self-learning and self-optimization capabilities, and enhances risk response capabilities.

CN123117976APending Publication Date: 2026-11-17ZHONGJING TENGYUAN SUPPLY CHAIN TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611027281.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-10
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing supply chain inventory management technologies suffer from problems such as data silos in heterogeneous systems, lack of a unified global view, reliance on experience or simple formulas for decision-making, difficulty in achieving Pareto optimality through multi-objective optimization, and lack of closed-loop feedback mechanisms.

Method used

A data fusion layer is used to fuse heterogeneous data, a digital twin model is built for simulation, an intelligent decision engine is used for multi-objective optimization, and a closed-loop feedback module is used to realize iterative updates of strategies, including semantic mapping, time alignment, multi-objective optimization solution and real-time risk warning.

Benefits of technology

It has achieved a globally consistent data foundation for supply chain inventory management, reduced the cost of strategy trial and error, achieved Pareto optimal balance in multi-objective optimization, and has self-learning and self-optimization capabilities, thereby improving risk response capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117976A_ABST
    Figure CN123117976A_ABST
Patent Text Reader

Abstract

This invention provides a service platform for precise supply chain inventory management, comprising: a data fusion layer for accessing multiple heterogeneous data sources and fusing them to generate a standardized fused inventory dataset; a digital twin engine for constructing a digital twin model of supply chain inventory flow and simulating and evaluating the performance of various candidate inventory management strategies; an intelligent decision engine for performing multi-objective optimization under operational constraints, using multiple conflicting inventory management indicators as optimization targets, to determine the optimal strategy; and a closed-loop feedback module for outputting the optimal strategy to the execution system, collecting actual operational data and feeding it back to the data fusion layer to trigger iterative updates of the model and decisions. This invention, through the combination of data fusion, digital twin simulation, and multi-objective optimization, achieves precise management and continuous optimization of supply chain inventory, effectively reducing inventory costs, improving turnover efficiency, and reducing the risk of stockouts.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of supply chain management and intelligent decision-making technology, specifically to a service platform for precise management of supply chain inventory. Background Technology

[0002] Supply chain inventory management is a core aspect of enterprise management, directly impacting operating costs, capital turnover efficiency, and customer satisfaction. With the acceleration of global economic integration and increasingly fierce market competition, supply chain structures are becoming more complex, with a large number of widely distributed inventory nodes, posing unprecedented challenges to inventory management.

[0003] However, existing technologies have the following shortcomings: First, the data formats, semantic definitions, and time granularities of heterogeneous systems such as ERP, WMS, and TMS in the supply chain vary, forming information silos and making it difficult to integrate them into a globally unified inventory view.

[0004] Second, existing decision-making relies heavily on experience or simple formulas, lacking dynamic simulation verification methods that cover the entire supply chain, and thus cannot comprehensively compare and evaluate multiple candidate strategies.

[0005] Third, inventory management naturally faces multiple conflicting optimization objectives such as cost, turnover rate, and stockout rate. Existing systems often use a single indicator or simple weighting to handle these objectives, making it difficult to obtain a Pareto optimal balance solution.

[0006] Fourth, most existing systems are open-loop, lacking feedback and model correction mechanisms for actual implementation effects after decision output, and lack the ability for continuous iterative optimization.

[0007] Therefore, this solution proposes a service platform for precise supply chain inventory management to address the aforementioned issues. Summary of the Invention

[0008] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a service platform for precise management of supply chain inventory.

[0009] To achieve the aforementioned objective, the technical solution of the present invention is implemented as follows: a service platform for precise supply chain inventory management, comprising: The data fusion layer is used to access multiple heterogeneous data sources in the supply chain, merge and process inventory management-related data from different data sources, and generate a standardized fused inventory dataset. The digital twin engine is used to build a digital twin model representing the inventory flow status of the supply chain based on a fused inventory dataset, and to simulate and run multiple candidate inventory management strategies in the digital twin model to obtain the performance evaluation results of each candidate strategy. An intelligent decision engine is used to determine the optimal inventory management strategy from multiple candidate strategies by performing multi-objective optimization under preset operational constraints, based on performance evaluation results and using at least two conflicting inventory management indicators as optimization objectives; and... The closed-loop feedback module is used to output the optimal inventory management strategy to the supply chain execution system and collect actual inventory operation data to feed back to the data fusion layer. The data fusion layer updates the fused inventory dataset based on the actual inventory operation data to trigger iterative updates of the digital twin engine and intelligent decision engine.

[0010] Preferably, the data fusion layer performs fusion processing on inventory management-related data from different data sources, including: semantic mapping of inventory-related fields in different data sources, association matching of records of the same inventory event in different data sources, and time alignment of time-series data at different time granularities.

[0011] Preferably, when the digital twin engine simulates multiple candidate inventory management strategies in the digital twin model, it also introduces supply chain disturbance factors for simulation and deduction. The performance evaluation results include the performance indicators of each candidate strategy under normal and disturbance scenarios.

[0012] Preferably, the optimization objectives of the intelligent decision engine for multi-objective optimization include minimizing the total inventory cost of the supply chain, maximizing the inventory turnover rate, and minimizing the stockout rate. The preset operating constraints include the storage capacity constraints of each inventory node, the supplier supply capacity constraints, and the transportation timeliness constraints.

[0013] Preferably, after collecting actual inventory operation data, the closed-loop feedback module calculates the deviation between actual performance and performance evaluation results. When the deviation exceeds a preset threshold, the model parameters of the digital twin model are corrected online.

[0014] Preferably, the service platform also includes an inventory risk early warning module, which is used to monitor the inventory status of each inventory node in real time. When the real-time inventory is lower than the safety stock level, a first-level early warning signal is generated. When the digital twin engine simulation shows that a shortage will occur within a preset future time period, a second-level early warning signal is generated.

[0015] Preferably, when the inventory risk warning module generates a second-level warning signal, it calls the intelligent decision engine to re-solve the multi-objective optimization problem with the current inventory status as the initial state, and outputs the optimal inventory management strategy obtained from the re-solved problem as an emergency replenishment plan.

[0016] Preferably, the service platform also includes a visualization module, which is used to display the simulation running status of the digital twin model and the decision basis of the optimal inventory management strategy output by the intelligent decision engine in the visualization interface.

[0017] Preferably, the merged inventory dataset includes at least real-time inventory data, in-transit inventory data, and historical demand data for each inventory node.

[0018] Preferably, the service platform is deployed using a microservice architecture, with the data fusion layer, digital twin engine, intelligent decision engine, and closed-loop feedback module running as independent microservices, and the microservices interacting with each other through asynchronous communication.

[0019] The beneficial effects of this invention are reflected in: This invention effectively solves the data silo problem between heterogeneous systems in the supply chain through semantic mapping, association matching, and time alignment processing in the data fusion layer. It generates a standardized, globally consistent fused inventory dataset, providing a reliable data foundation for precise inventory management. By constructing a digital twin engine covering the entire supply chain's inventory flow status, it can accurately map the physical supply chain's inventory dynamics in virtual space and support simulation and performance evaluation of multiple candidate strategies, significantly reducing the cost of strategy trial and error. Furthermore, through an intelligent decision engine, it performs multi-objective optimization with multiple conflicting inventory management indicators as optimization targets, enabling… It achieves Pareto optimal balance among multiple dimensions such as inventory cost, turnover rate, and stockout rate, overcoming the limitations of traditional single-indicator optimization or simple weighting methods. Through a closed-loop feedback module, it realizes a complete closed loop from strategy output to actual execution, data feedback, and model updates, enabling the platform to have self-learning and self-optimization capabilities and continuously improve decision-making accuracy based on actual operating conditions. Through an inventory risk early warning module, it realizes real-time monitoring of inventory status and forward-looking risk assessment. Combined with the dual mechanism of first-level real-time early warning and second-level forward-looking early warning, as well as the automatic generation of emergency replenishment plans, it significantly improves the risk response capability of supply chain inventory management. Attached Figure Description

[0020] In the attached diagram: Figure 1 This is a schematic diagram of the system architecture of the service platform for precise supply chain inventory management according to the present invention; Figure 2 This is a schematic diagram of the data fusion processing flow of the data fusion layer of the present invention; Figure 3 This is a schematic diagram of the simulation operation process of the digital twin engine of the present invention; Figure 4 This is a schematic diagram of the multi-objective optimization solution process of the intelligent decision engine of the present invention; Figure 5 This is a schematic diagram of the iterative update process of the closed-loop feedback module of the present invention; Figure 6 This is a schematic diagram of the early warning process of the inventory risk early warning module of the present invention. Detailed Implementation

[0021] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the invention, and not all of them. Unless otherwise specified, the embodiments and features described in this application can be combined with each other. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0022] It should be noted that if the embodiments of the invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0023] Furthermore, "multiple" refers to two or more. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of a person skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the invention.

[0024] Example 1: System Overall Architecture like Figure 1 As shown, the service platform for precise supply chain inventory management provided by this invention includes a data fusion layer, a digital twin engine, an intelligent decision engine, a closed-loop feedback module, an inventory risk early warning module, and a visualization module. The service platform is deployed using a microservice architecture, with the data fusion layer, digital twin engine, intelligent decision engine, and closed-loop feedback module each running as an independent microservice. Data interaction between these microservices occurs asynchronously.

[0025] Example 2: Detailed Implementation of the Data Fusion Layer The data fusion layer is used to access multiple heterogeneous data sources in the supply chain (including ERP, WMS, TMS, supplier collaboration platforms, etc.) and perform the following fusion processes (such as...). Figure 2 (as shown) Step S201: Semantic mapping – Map inventory-related fields (such as “material number”, “product code”, and “SKU number”) from various heterogeneous data sources to the same field in the standard data model.

[0026] Step S202: Event Association Matching – By using key fields such as order number, timestamp, product code, and quantity, records of the same inventory event (such as purchase receipt) in different systems are associated as a single event to avoid duplication or omission.

[0027] Step S203: Time alignment – ​​Use interpolation or aggregation methods to unify time series data from different data sources with different time granularities (such as minute-level, hour-level, and day-level) to the same time granularity.

[0028] After the above processing, a standardized fused inventory dataset is generated, which includes at least the real-time inventory, in-transit inventory, and historical demand data for each inventory node.

[0029] Example 3: Detailed Implementation of the Digital Twin Engine The digital twin engine builds a digital twin model of the supply chain inventory flow status based on a fused inventory dataset (such as...). Figure 3 As shown in the figure, its workflow is as follows: Step S301: Model Building - Establish inventory dynamic equations for each inventory node, supply relationship models between nodes, and end-demand models, covering the entire chain nodes such as raw materials, work-in-process, finished products, distribution centers, and retailers.

[0030] Step S302: Candidate Strategy Configuration - Receive multiple candidate inventory management strategies (including EOQ strategy with different parameters, (s,S) strategy, periodic inventory count strategy, dynamic replenishment strategy, etc.), each strategy is input in the form of configurable parameters.

[0031] Step S303: Simulation run – Simulate the dynamic changes in inventory under various strategies within a preset time span using discrete events or system dynamics.

[0032] Step S304: Introduction of Disturbance Scenarios - Overlay supply chain disturbance factors (including supplier delayed delivery, transportation disruption, demand fluctuations, capacity constraints, etc.) to conduct simulations of both normal and disturbance scenarios.

[0033] Step S305: Performance Evaluation – Calculate the performance indicators of each strategy in the dual scenarios, including total inventory cost (holding cost + ordering cost + stockout cost), inventory turnover rate, stockout rate, average inventory level, and service level.

[0034] Example 4: Detailed Implementation of the Intelligent Decision Engine The intelligent decision engine performs multi-objective optimization solutions based on performance evaluation results (such as...). Figure 4 (as shown) Step S401: Set optimization objectives – minimize total supply chain inventory cost, maximize inventory turnover rate, and minimize stockout rate as three conflicting optimization objectives.

[0035] Step S402: Set operational constraints – including storage capacity constraints for each inventory node (inventory quantity ≤ maximum capacity), supplier supply capacity constraints (single replenishment quantity ≤ capacity limit), and transportation timeliness constraints (replenishment lead time ≥ minimum transportation time).

[0036] Step S403: Multi-objective optimization solution – Use Pareto optimality algorithms (such as NSGA-II, MOEA / D, or reinforcement learning-based methods) to solve for the Pareto optimal solution set from candidate policies.

[0037] Step S404: Determine the optimal strategy – Select the final optimal strategy from the Pareto optimal solution set based on the preset decision preferences (weights or interactive selection).

[0038] Step S405: Generate decision basis - output the target values ​​of the optimal strategy, the constraint satisfaction status, and a comparative analysis with the suboptimal strategy.

[0039] Example 5: Detailed Implementation of the Closed-Loop Feedback Module The closed-loop feedback module realizes strategy output, data acquisition and iterative updates (such as...). Figure 5 (as shown) Step S501: Strategy Output - Output the optimal strategy as standardized instructions (replenishment suggestions, inventory adjustment notifications) to the supply chain execution system (WMS, procurement system, TMS, etc.).

[0040] Step S502: Actual Data Collection - Continuously collect operational data such as actual inventory levels, replenishment execution, actual demand, and actual stockouts after execution.

[0041] Step S503: Deviation Calculation – Compare the actual performance with the simulated predicted performance and calculate the absolute and relative deviations of each indicator.

[0042] Step S504: Threshold Judgment and Model Correction - When the deviation exceeds the preset threshold, the parameters of the digital twin model (such as demand forecast coefficient, replenishment lead time, supplier capability parameters, etc.) are automatically adjusted.

[0043] Step S505: Data Feedback and Iteration Triggering – The actual running data is sent back to the data fusion layer to update the fused dataset, thereby triggering the reconstruction of the digital twin model and the re-optimization of the intelligent decision engine, forming a continuous iterative closed loop.

[0044] Example 6: Detailed Implementation of the Inventory Risk Early Warning Module The inventory risk warning module monitors the inventory status at each node in real time (e.g., Figure 6 (as shown) Step S601: Real-time monitoring – Obtain the real-time inventory of each node through the data fusion layer.

[0045] Step S602: Level 1 Warning - When the real-time inventory is lower than the preset safety stock level, a Level 1 warning signal is immediately generated, which includes the node identifier, current inventory, safety stock level, and suggested replenishment quantity.

[0046] Step S603: Second-level early warning - The digital twin engine starts from the current state and simulates the inventory changes over a future preset time period (such as 7 days or 30 days). If a stockout is predicted to occur, a second-level forward-looking early warning signal is generated.

[0047] Step S604: Emergency Replenishment Plan - When a Level 2 alert is generated, the intelligent decision engine is automatically invoked to re-optimize the multi-objectives based on the current inventory as the initial state, and outputs an emergency replenishment plan (including replenishment nodes, quantities, timing, and suppliers).

[0048] Example 7: Detailed Implementation of the Visualization Module The visualization module provides the following display functions in the interface: Digital twin visualization: graphically displaying real-time inventory levels, inventory flow, and supply relationships at each node; Simulation process visualization: Displays the dynamic inventory change curve and replenishment / out-of-stock event sequence during simulation; Visualization of decision results: Displaying the distribution of Pareto optimal solution set, performance comparison of each strategy, and optimal strategy parameters; Decision justification presentation: Show the detailed decision justification for the optimal strategy, enhancing interpretability.

[0049] Example 8: Application Scenarios Taking a manufacturing company's multi-level supply chain (including 1 central warehouse, 3 regional distribution centers, 50 retail terminals, and over 2000 SKUs) as an example, after deploying this platform: The data fusion layer integrates with ERP, WMS, TMS, and supplier systems, completing semantic mapping, event correlation, and time alignment to generate a fused dataset. The digital twin engine constructs a 54-node digital twin model, generating multiple candidate strategies for each SKU (e.g., 5 sets of EOQ parameters and 3 sets of (s,S) parameters), simulating one year of performance under normal and disruptive scenarios (delayed delivery, sudden demand surges). The intelligent decision engine, targeting cost, turnover rate, and stockout rate, uses NSGA-II to solve for Pareto optimal strategies under constraints of warehouse capacity, supply capacity, and transportation timeliness, selecting the final strategy according to enterprise preferences. The closed-loop feedback module outputs the strategy to the execution system, collecting actual data in real time. When the turnover rate deviation exceeds 10%, it automatically corrects the model parameters and re-optimizes. The risk warning module monitors inventory in real time, triggering a first-level warning when it falls below the safety level, and a second-level warning when a stockout is predicted for the next two weeks, automatically generating an emergency replenishment plan. All states and decision results are presented through a visual interface.

[0050] 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 within the protection scope of the present invention.

[0051] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0052] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A service platform for precise supply chain inventory management, characterized in that, include: The data fusion layer is used to access multiple heterogeneous data sources in the supply chain, merge and process inventory management-related data from different data sources, and generate a standardized fused inventory dataset. A digital twin engine is used to construct a digital twin model representing the inventory flow status of the supply chain based on the fused inventory dataset, and to simulate and run multiple candidate inventory management strategies in the digital twin model to obtain the performance evaluation results of each candidate strategy. The intelligent decision engine is used to perform multi-objective optimization under preset operating constraints based on the performance evaluation results, with at least two conflicting inventory management indicators as optimization targets, to determine the optimal inventory management strategy from the multiple candidate inventory management strategies. as well as, The closed-loop feedback module is used to output the optimal inventory management strategy to the supply chain execution system and collect actual inventory operation data to feed back to the data fusion layer. The data fusion layer updates the fused inventory dataset according to the actual inventory operation data to trigger the iterative update of the digital twin engine and the intelligent decision engine.

2. The service platform for precision management of supply chain inventory according to claim 1, wherein, The data fusion layer performs fusion processing on inventory management-related data from different data sources, including: semantic mapping of inventory-related fields in different data sources, association matching of records of the same inventory event in different data sources, and time alignment of time-series data at different time granularities.

3. The service platform for precision management of supply chain inventory according to claim 1, wherein, The digital twin engine simulates various candidate inventory management strategies in the digital twin model, and also introduces supply chain disturbance factors for simulation. The performance evaluation results include the performance indicators of each candidate strategy under normal and disturbance scenarios.

4. The service platform for precision management of supply chain inventory according to claim 1, wherein, The optimization objectives of the intelligent decision engine for multi-objective optimization include minimizing the total inventory cost of the supply chain, maximizing the inventory turnover rate, and minimizing the stockout rate. The preset operating constraints include the storage capacity constraints of each inventory node, the supplier supply capacity constraints, and the transportation timeliness constraints.

5. The service platform for precision management of supply chain inventory according to claim 1, wherein, After collecting actual inventory operation data, the closed-loop feedback module calculates the deviation between the actual performance and the performance evaluation result. When the deviation exceeds a preset threshold, the model parameters of the digital twin model are corrected online.

6. The service platform for precision management of supply chain inventory according to claim 1, wherein, It also includes an inventory risk early warning module, which is used to monitor the inventory status of each inventory node in real time. When the real-time inventory is lower than the safety stock level, a first-level early warning signal is generated. When the simulation and deduction of the digital twin engine indicates that a shortage will occur within a preset future time period, a second-level early warning signal is generated.

7. The service platform for supply chain inventory precision management of claim 6, wherein, When the inventory risk early warning module generates the second-level early warning signal, it calls the intelligent decision engine to re-solve the multi-objective optimization problem with the current inventory status as the initial state, and outputs the optimal inventory management strategy obtained from the re-solved problem as an emergency replenishment plan.

8. The service platform for precision management of supply chain inventory according to claim 1, wherein, It also includes a visualization module, which is used to display the simulation running status of the digital twin model and the decision basis of the optimal inventory management strategy output by the intelligent decision engine in a visualization interface.

9. The service platform for precision management of supply chain inventory according to claim 1, wherein, The fused inventory dataset includes at least real-time inventory data, in-transit inventory data, and historical demand data for each inventory node.

10. The service platform for precise supply chain inventory management according to claim 1, characterized in that, The service platform is deployed using a microservice architecture. The data fusion layer, the digital twin engine, the intelligent decision engine, and the closed-loop feedback module each run as independent microservices, and the microservices interact with each other through asynchronous communication.