Smart operation method and system of commercial complex based on digital twin and IOT

Abstract

The present invention discloses a commercial complex intelligent operation method and system based on digital twin and IOT, which relates to the field of commercial complex operation technology, including: when inventory changes are detected, data is transmitted to a cloud data center in real time; key feature information of the equipment is extracted; an intelligent inventory model of the equipment is constructed and dynamically updated, and matching is performed in the model according to the extracted feature information to generate a corresponding inventory forecasting scheme; according to the generated inventory forecasting scheme, the inventory management task is automatically divided into multiple parallel subtasks; based on the matching result, whether the conditions for automatic inventory adjustment are met is evaluated, and if so, the adjustment is automatically performed, and if not, the adjustment is performed after confirmation by the inventory manager; and future inventory demand is predicted in combination with historical inventory data and usage frequency. The present invention helps to timely discover abnormal inventory conditions, reduce decision-making errors caused by lagging inventory information, improve the real-time and accuracy of inventory management, and optimize the inventory structure.

Description

Technical Field

[0001] The present invention relates to the field of commercial complex operation technology, and in particular to a commercial complex intelligent operation method and system based on digital twin and IOT. Background Art

[0002] With the rapid development of information technology, digital twin technology and the Internet of Things (IoT) have become important means to promote the intelligent management of commercial complexes. Digital twin technology simulates system behavior in the real world by creating virtual mappings of physical entities, while IoT connects physical objects to the Internet to achieve communication and data exchange between devices. In recent years, these two technologies have been widely used in the fields of building management and operation, providing efficient data collection and analysis capabilities for commercial complexes. Digital twin technology can achieve all-round monitoring and management of buildings and their facilities by integrating multi-source heterogeneous data, while IoT sensor networks can collect environmental and equipment data in real time, providing basic support for refined management. In addition, the development of cloud computing platforms has also greatly promoted the ability of data processing and storage, enabling commercial complexes to better utilize big data resources for decision support.

[0003] Although digital twin technology and IoT have played an important role in the operation and management of commercial complexes, there are still some limitations. On the one hand, traditional inventory management systems often rely on manual regular inspections and records, which is not only inefficient but also prone to human errors; on the other hand, the lack of effective data analysis methods makes it impossible to fully utilize real-time data for accurate predictions and optimized decisions. In addition, the current inventory management system usually only focuses on the current status and ignores historical trend analysis, making it difficult to accurately predict future demand changes, which in turn affects the formulation of procurement plans. Therefore, in the operation and management of commercial complexes, how to effectively integrate real-time data streams and realize intelligent inventory management has become an urgent problem to be solved. The present invention aims to solve the above-mentioned problems existing in the prior art by constructing an intelligent inventory management system based on digital twins and IoT, and realize real-time monitoring, intelligent analysis and automatic adjustment of inventory. Summary of the invention

[0004] In view of the problems existing in the existing commercial complex intelligent operation method based on digital twin and IOT, the present invention is proposed. Therefore, the problem to be solved by the present invention is how to provide a commercial complex intelligent operation method and system based on digital twin and IOT.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In the first aspect, the present invention provides a smart operation method for a commercial complex based on digital twins and IOT, which includes: using an Internet of Things sensor network to monitor the inventory status of equipment in real time, and when inventory changes are detected, transmitting data to a cloud data center in real time; executing an intelligent classification program on the data received from the cloud to extract key feature information of the equipment; building and dynamically updating an intelligent inventory model for the equipment, matching the extracted feature information in the model to generate a corresponding inventory forecasting plan; automatically dividing the inventory management task into multiple parallel subtasks according to the generated inventory forecasting plan; evaluating whether the conditions for automated inventory adjustment are met based on the matching results, and automatically executing the adjustment if the conditions are met, and executing the adjustment after confirmation by the inventory administrator if the conditions are not met; applying a time series prediction algorithm, combining historical inventory data and usage frequency to predict future inventory demand, and providing data support for procurement decisions.

[0007] As a preferred solution of the intelligent operation method of a commercial complex based on digital twins and IOT described in the present invention, the intelligent classification program for the data received from the cloud includes the following steps: when executing the intelligent classification program, an adaptive feature extraction mechanism based on deep learning is introduced; based on the device domain knowledge, a feature extraction rule set E is constructed; for each cloud data D, a multi-dimensional analysis is performed based on the feature extraction rule set E to obtain a feature vector F; a feature quality evaluation function Q(F, E) is defined to evaluate the accuracy of the feature vector F under rule E, and a threshold parameter β is set; if Q(F, E) ≥ β, F is used as the feature representation of the device; if Q(F, E) < β, a manual review mechanism is triggered, and the feature extraction rule E is optimized by an expert; based on the rule E' optimized by the expert, the feature quality evaluation function Q'(F, E') is reconstructed; and the cloud data D is re-extracted using the updated evaluation function Q'(F, E').

[0008] As a preferred solution of the commercial complex intelligent operation method based on digital twin and IOT described in the present invention, the specific formula of the feature quality evaluation function is:

[0009] Q(F,E)=γ1C(F)+γ2I(F,E)+γ3R(F,H)C(F)

[0010]

[0011] I(F,E)=∑ e∈E δ(e,F) / |E|

[0012] R(F,H)=exp(-D KL (P F |P H ))

[0013] Among them, C(F) represents the independence measure of feature vector F, I(F,E) represents the coverage of F to rule E, R(F,H) represents the correlation between F and historical data H, γ1, γ2, and γ3 are weight parameters; corr(f i ,f j ) represents the feature f i and f j The correlation coefficient of is, n is the number of features, δ(e,F) represents the applicability of rule e in F, |E| is the size of the rule set, D KL represents KL divergence, P F and P H Denote the probability distribution of F and H respectively; the specific formula for reconstructing the feature quality evaluation function Q'(F, E') is:

[0014] Q′(F,E′)=(1-ω)Q(F,E′)+ω·S(F,E′)

[0015] Among them, S(F,E') represents the expert score and ω is the balance parameter.

[0016] As a preferred solution of the smart operation method of a commercial complex based on digital twins and IOT described in the present invention, the construction and dynamic updating of the equipment intelligent inventory model includes the following steps: based on historical inventory data and usage records, deep learning technology is used to automatically extract equipment usage patterns and inventory change rules; multi-source heterogeneous data from supply chain systems, logistics information systems and equipment management systems are integrated to build a comprehensive equipment life cycle database; the equipment inventory management problem is modeled as a dynamic time series prediction task, and a recurrent neural network model with an attention mechanism is introduced for inventory prediction; inventory management knowledge is shared and transferred between different types of equipment using transfer learning technology; for newly introduced equipment, a few-sample learning method is used to construct an initial inventory model; model parameters are continuously optimized through an online learning algorithm, and personalized and explainable inventory management recommendations are dynamically generated based on the prediction results and actual inventory conditions.

[0017] As a preferred solution of the commercial complex intelligent operation method based on digital twin and IOT described in the present invention, the integration of multi-source heterogeneous data from the supply chain system, logistics information system and equipment management system includes the following steps: designing a heterogeneous data fusion framework F including three subsystems of supply chain SC, logistics information HI and equipment management EM; in each subsystem S i Train the autoencoder model separately to obtain the low-dimensional representation vector v i ; Define the cross-system data alignment function, the expression is:

[0018] A(v i ,v j )=tanh(W·[vi ;v j ]+b)

[0019] Where W is the weight matrix and b is the bias vector; construct the global alignment optimization objective, the expression is:

[0020]

[0021] Among them, K i,j represents the set of known aligned data pairs in systems i and j, and λ is the regularization parameter. L is optimized by the gradient descent method to obtain the optimal cross-system data alignment solution. Based on the alignment results, a unified device data representation is constructed to achieve the fusion of multi-source heterogeneous data.

[0022] As a preferred solution of the commercial complex intelligent operation method based on digital twin and IOT described in the present invention, the automatic division of the inventory management task into multiple parallel subtasks includes the following steps: constructing a multi-dimensional task decomposition strategy based on the attributes of the equipment and the inventory characteristics; formalizing the inventory management problem into a multi-objective optimization problem, while considering the minimization of inventory costs and the maximization of service levels; applying a genetic algorithm to generate an initial set of task decomposition solutions, and designing a fitness function, which is expressed as follows:

[0023] F(x)=α·Cost(x)+β·ServiceLevel(x)

[0024] Among them, x represents the task decomposition scheme, α and β are weight coefficients; the task decomposition scheme is iteratively optimized through crossover, mutation and selection operations until convergence or the maximum number of iterations is reached; for the generated optimal task decomposition scheme, parallel computing technology is applied to simultaneously execute multiple subtasks; a dynamic adjustment mechanism is set up to adaptively adjust the task decomposition strategy according to the real-time inventory status and execution effect.

[0025] As a preferred solution of the commercial complex intelligent operation method based on digital twin and IOT described in the present invention, the setting of the dynamic adjustment mechanism includes the following steps: defining the state space S as the current inventory level and task execution progress, the action space A as the possible task adjustment operations, and constructing a reward function, which is expressed as:

[0026] R(s,a,s′)=w1ΔCost+w2ΔServiceLevel+w3ΔEfficiency

[0027] Among them, s and s' represent the states before and after adjustment, ΔCost, ΔServiceLevel and ΔEfficiency represent the changes in cost, service level and efficiency, respectively, and w1, w2 and w3 are weights. The deep Q learning algorithm is used to learn the optimal dynamic adjustment strategy through experience replay and target network technology. The loss function of the deep Q learning algorithm is defined as,

[0028] L(θ)=E[(R+γmax′ a P(s′,a′;θ′)-P(s,a;θ)) 2 ]

[0029] Among them, θ and θ' represent the parameters of the current network and the target network respectively, and γ is the discount factor. During the strategy execution process, a greedy strategy is used to balance exploration and utilization, and the strategy performance is evaluated regularly. If the performance deteriorates, the strategy retraining mechanism is triggered.

[0030] In the second aspect, the present invention provides a smart operation system for a commercial complex based on digital twins and IOT, which includes: a detection module, which is used to monitor the inventory status of equipment in real time using an Internet of Things sensor network, and when inventory changes are detected, the data is transmitted to the cloud data center in real time; an extraction module, which is used to execute an intelligent classification program on the data received from the cloud to extract key feature information of the equipment; a construction module, which is used to build and dynamically update the intelligent inventory model of the equipment, match the extracted feature information in the model, and generate a corresponding inventory forecasting plan; a division module, which is used to automatically divide the inventory management task into multiple parallel subtasks according to the generated inventory forecasting plan; an adjustment module, which is used to evaluate whether the conditions for automated inventory adjustment are met based on the matching results, and if so, automatically perform the adjustment, and if not, execute it after confirmation by the inventory administrator; a prediction module, which is used to apply a time series prediction algorithm, combine historical inventory data and usage frequency, predict future inventory demand, and provide data support for procurement decisions.

[0031] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, it implements the steps of a smart operation method for a commercial complex based on digital twins and IOT.

[0032] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, the steps of a method for intelligent operation of a commercial complex based on digital twins and IOT are implemented.

[0033] The beneficial effects of the present invention are that it helps to timely discover inventory anomalies, can respond quickly, reduce decision-making errors caused by lagging inventory information, improve the real-time and accuracy of inventory management, enhance the intelligence level of data processing, reduce inventory costs and risks, enhance the flexibility and response speed of inventory management, provide a scientific basis for procurement decisions, and optimize inventory structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 This is a flow chart of the smart operation method of a commercial complex based on digital twins and IOT. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned purposes, features and advantages of the present invention more understandable, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0037] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0038] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0039] Example 1

[0040] Reference Figure 1 , which is the first embodiment of the present invention, and provides a commercial complex intelligent operation method based on digital twins and IOT, including:

[0041] S1: Use IoT sensor networks to monitor equipment inventory status in real time, and when inventory changes are detected, transmit data to the cloud data center in real time;

[0042] S2: Execute intelligent classification procedures on the data received from the cloud to extract key feature information of the device;

[0043] When executing the intelligent classification procedure, an adaptive feature extraction mechanism based on deep learning is introduced;

[0044] Based on the equipment domain knowledge, construct the feature extraction rule set E;

[0045] For each piece of cloud data D, perform multi-dimensional analysis based on the feature extraction rule set E to obtain the feature vector F;

[0046] Define the feature quality evaluation function Q(F,E) to evaluate the accuracy of the feature vector F under rule E, and set the threshold parameter β;

[0047] The specific formula of the feature quality evaluation function is:

[0048] Q(F,E)=γ1C(F)+γ2I(F,E)+γ3R(F,H)C(F)

[0049]

[0050] I(F,E)=∑ e∈E δ(e,F) / |E|

[0051] R(F,H)=exp(-D KL (P F |P H ))

[0052] Among them, C(F) represents the independence measure of feature vector F, I(F,E) represents the coverage of F to rule E, R(F,H) represents the correlation between F and historical data H, γ1, γ2, and γ3 are weight parameters; corr(f i ,f j ) represents the feature f i and f j , n is the number of features, δ(e,F) represents the applicability of rule e in F, |E| is the size of the rule set, D KL represents KL divergence, P F and P H Represent the probability distribution of F and H respectively;

[0053] If Q(F,E)≥β, F is used as the characteristic representation of the device;

[0054] If Q(F,E)<β, the manual review mechanism is triggered, and the expert optimizes the feature extraction rule E;

[0055] Based on the expert-optimized rule E', the feature quality evaluation function Q'(F, E') is reconstructed; the specific formula for reconstructing the feature quality evaluation function Q'(F, E') is:

[0056] Q′(F,E′)=(1-ω)Q(F,E′)+ω·S(F,E′)

[0057] Among them, S(F,E') represents the expert score and ω is the balance parameter.

[0058] Use the updated evaluation function Q'(F,E') to re-extract features from the cloud data D.

[0059] S3: Build and dynamically update the equipment intelligent inventory model, match the extracted feature information in the model, and generate the corresponding inventory forecast plan;

[0060] Based on historical inventory data and usage records, deep learning technology is used to automatically extract equipment usage patterns and inventory change patterns;

[0061] Integrate multi-source heterogeneous data from supply chain systems, logistics information systems, and equipment management systems to build a comprehensive equipment lifecycle database;

[0062] Design a heterogeneous data fusion framework F that includes three subsystems: supply chain SC, logistics information HI, and equipment management EM;

[0063] In each subsystem S i Train the autoencoder model separately to obtain the low-dimensional representation vector v i ;

[0064] Define the cross-system data alignment function, the expression is:

[0065] A(v i ,v j )=tanh(W·[v i ;v j ]+b)

[0066] Where W is the weight matrix and b is the bias vector;

[0067] Construct the global alignment optimization objective, the expression is:

[0068]

[0069] Among them, K i,j represents the set of known aligned data pairs in systems i and j, and λ is the regularization parameter;

[0070] Optimize L through the gradient descent method to obtain the optimal cross-system data alignment solution;

[0071] Based on the alignment results, a unified device data representation is constructed to achieve the fusion of multi-source heterogeneous data;

[0072] The equipment inventory management problem is modeled as a dynamic time series prediction task, and a recurrent neural network model with attention mechanism is introduced for inventory prediction.

[0073] Use transfer learning technology to share and transfer inventory management knowledge between different types of equipment to improve the generalization ability of the model;

[0074] For newly introduced devices, a few-shot learning method is used to quickly build an initial inventory model;

[0075] Continuously optimize model parameters through online learning algorithms to adapt to dynamic changes in device usage patterns;

[0076] Combine the forecast results with the actual inventory situation to dynamically generate personalized and explainable inventory management recommendations.

[0077] S4: According to the generated inventory forecasting plan, the inventory management task is automatically divided into multiple parallel subtasks;

[0078] Build a multi-dimensional task decomposition strategy based on equipment attributes and inventory characteristics;

[0079] The inventory management problem is formalized as a multi-objective optimization problem, which considers the minimization of inventory cost and the maximization of service level at the same time;

[0080] Genetic algorithm is used to generate the initial set of task decomposition solutions and the fitness function is designed. The expression is:

[0081] F(x)=α·Cost(x)+β·ServiceLevel(x)

[0082] Among them, x represents the task decomposition scheme, α and β are weight coefficients;

[0083] Iteratively optimize the task decomposition scheme through crossover, mutation and selection operations until convergence or the maximum number of iterations is reached;

[0084] For the generated optimal task decomposition solution, parallel computing technology is applied to execute multiple subtasks simultaneously;

[0085] Set up a dynamic adjustment mechanism to adaptively adjust the task decomposition strategy based on real-time inventory status and execution results.

[0086] Define the state space S as the current inventory level and task execution progress, the action space A as possible task adjustment operations, and construct the reward function, which is expressed as:

[0087] R(s,a,s′)=w1ΔCost+w2ΔServiceLevel+w3ΔEfficiency

[0088] Among them, s and s' represent the status before and after adjustment, ΔCost, ΔServiceLevel and ΔEfficiency represent the changes in cost, service level and efficiency, respectively, and w1, w2 and w3 are weights;

[0089] The deep Q learning algorithm is used to learn the optimal dynamic adjustment strategy through experience replay and target network technology. The loss function of the deep Q learning algorithm is defined as:

[0090] L(θ)=E[(R+γmax′ a P(s′,a′;θ′)-Q(s,a;P)) 2 ]

[0091] Among them, θ and θ' represent the parameters of the current network and the target network respectively, and γ is the discount factor

[0092] During the strategy execution process, a greedy strategy is used to balance exploration and utilization, and the strategy performance is evaluated regularly. If the performance degrades, the strategy retraining mechanism is triggered.

[0093] S5: Based on the matching results, evaluate whether the conditions for automatic inventory adjustment are met. If so, the adjustment is automatically performed. If not, the adjustment is performed after confirmation by the inventory manager.

[0094] S6: Apply time series forecasting algorithms, combined with historical inventory data and usage frequency, to predict future inventory demand and provide data support for purchasing decisions.

[0095] Furthermore, the present embodiment also provides a smart operation system for a commercial complex based on digital twins and IOT, including: a detection module, which is used to monitor the inventory status of equipment in real time using an Internet of Things sensor network, and when inventory changes are detected, the data is transmitted to the cloud data center in real time; an extraction module, which is used to execute an intelligent classification program on the data received from the cloud to extract key feature information of the equipment; a construction module, which is used to build and dynamically update the intelligent inventory model of the equipment, match the extracted feature information in the model, and generate a corresponding inventory forecasting plan; a division module, which is used to automatically divide the inventory management task into multiple parallel subtasks according to the generated inventory forecasting plan; an adjustment module, which is used to evaluate whether the conditions for automated inventory adjustment are met based on the matching results, and if so, automatically perform the adjustment, and if not, execute it after confirmation by the inventory administrator; a prediction module, which is used to apply a time series prediction algorithm, combine historical inventory data and usage frequency, predict future inventory demand, and provide data support for procurement decisions.

[0096] This embodiment also provides a computer device, which is suitable for the intelligent operation method of a commercial complex based on digital twins and IOT, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement all or part of the steps of the method described in the embodiment of the present invention as proposed in the above embodiment.

[0097] This embodiment also provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method in any optional implementation of the above embodiment is executed. Wherein, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, disk or optical disk.

[0098] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0099] As can be seen from the above, this method realizes comprehensive and real-time monitoring of the inventory status of equipment, ensuring the accuracy and timeliness of inventory data. It helps to discover abnormal inventory situations in time, respond quickly, reduce decision-making errors caused by lagging inventory information, improve the real-time and accuracy of inventory management, enhance the intelligent level of data processing, and continuously optimize inventory management strategies through continuous learning and updating of models, avoid excessive inventory or out-of-stock situations, and reduce inventory costs and risks. It improves the execution efficiency of inventory management tasks, reduces labor costs, realizes a high degree of automation of inventory management, reduces the necessity of human intervention, and enhances the flexibility and response speed of inventory management. By comprehensively analyzing historical data and current usage, scientific predictions can be made for future demand trends, thereby guiding procurement activities, avoiding the risk of inventory backlogs or shortages, providing a scientific basis for procurement decisions, and optimizing inventory structure.

[0100] Example 2

[0101] Referring to Table 1, which is the second embodiment of the present invention, this embodiment provides a smart operation method for a commercial complex based on digital twins and IOT. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0102] Table 1 Test data table

[0103]

[0104] As can be seen from the above, this method realizes comprehensive and real-time monitoring of the inventory status of equipment, ensuring the accuracy and timeliness of inventory data. It helps to discover abnormal inventory situations in time, respond quickly, reduce decision-making errors caused by lagging inventory information, improve the real-time and accuracy of inventory management, enhance the intelligent level of data processing, and continuously optimize inventory management strategies through continuous learning and updating of models, avoid excessive inventory or out-of-stock situations, and reduce inventory costs and risks. It improves the execution efficiency of inventory management tasks, reduces labor costs, realizes a high degree of automation of inventory management, reduces the necessity of human intervention, and enhances the flexibility and response speed of inventory management. By comprehensively analyzing historical data and current usage, scientific predictions can be made for future demand trends, thereby guiding procurement activities, avoiding the risk of inventory backlogs or shortages, providing a scientific basis for procurement decisions, and optimizing inventory structure.

[0105] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A smart operation method for commercial complexes based on digital twins and IOT, characterized by: include, Use IoT sensor networks to monitor equipment inventory status in real time, and when inventory changes are detected, transmit data to the cloud data center in real time; Perform intelligent classification procedures on the data received from the cloud to extract key feature information of the device; Build and dynamically update the intelligent inventory model of equipment, match it in the model according to the extracted feature information, and generate the corresponding inventory forecast plan; According to the generated inventory forecast plan, the inventory management task is automatically divided into multiple parallel subtasks; Based on the matching results, the system evaluates whether the conditions for automatic inventory adjustment are met. If so, the adjustment is automatically performed. If not, the adjustment is performed after confirmation by the inventory manager. Apply time series forecasting algorithms, combine historical inventory data and usage frequency, and forecast future inventory demand to provide data support for purchasing decisions; The intelligent classification process for the data received from the cloud comprises the following steps: When executing the intelligent classification procedure, an adaptive feature extraction mechanism based on deep learning is introduced; Based on the equipment domain knowledge, construct the feature extraction rule set E; For each piece of cloud data D, perform multi-dimensional analysis based on the feature extraction rule set E to obtain the feature vector F; Define the feature quality evaluation function Q(F,E) to evaluate the accuracy of the feature vector F under rule E, and set the threshold parameter β; If Q(F,E)≥β, F is used as the characteristic representation of the device; If Q(F,E)<β, the manual review mechanism is triggered, and the expert optimizes the feature extraction rule E; Based on the expert-optimized rule E', the feature quality evaluation function Q'(F,E') is reconstructed; Use the updated evaluation function Q'(F,E') to re-extract features from the cloud data D; The specific formula of the feature quality evaluation function is: Q(F,E)=γ1C(F)+γ2I(F,E)+γ3R(F,H)C(F) I(F,E)=∑ e∈E δ(e,F) / |E| R(F,H)=exp(-D KL (P F |P H )) Among them, C(F) represents the independence measure of feature vector F, I(F,E) represents the coverage of F to rule E, R(F,H) represents the correlation between F and historical data H, γ1, γ2, and γ3 are weight parameters; corr(f i ,f j ) represents the feature f i and f j , n is the number of features, δ(e,F) represents the applicability of rule e in F, |E| is the size of the rule set, D KL represents KL divergence, P F and P H Represent the probability distribution of F and H respectively; The specific formula for reconstructing the feature quality evaluation function Q'(F, E') is: Q′(F,E′)=(1-ω)Q(F,E′)+ω·S(F,E′) Among them, S(F,E') represents the expert score and ω is the balance parameter.

2. The intelligent operation method of a commercial complex based on digital twins and IOT as claimed in claim 1, characterized in that: The construction and dynamic updating of the equipment intelligent inventory model includes the following steps: Based on historical inventory data and usage records, deep learning technology is used to automatically extract equipment usage patterns and inventory change patterns; Integrate multi-source heterogeneous data from supply chain systems, logistics information systems, and equipment management systems to build a comprehensive equipment lifecycle database; The equipment inventory management problem is modeled as a dynamic time series prediction task, and a recurrent neural network model with attention mechanism is introduced for inventory prediction. Use transfer learning technology to share and transfer inventory management knowledge between different types of equipment; For newly introduced devices, the few-shot learning method is used to build the initial inventory model; Through the online learning algorithm, the model parameters are continuously optimized, and the prediction results and actual inventory conditions are combined to dynamically generate personalized and explainable inventory management recommendations.

3. The intelligent operation method of a commercial complex based on digital twins and IOT as claimed in claim 2, characterized in that: The integration of multi-source heterogeneous data from the supply chain system, logistics information system and equipment management system includes the following steps: Design a heterogeneous data fusion framework F that includes three subsystems: supply chain SC, logistics information HI, and equipment management EM; In each subsystem S i Train the autoencoder model separately to obtain the low-dimensional representation vector v i ; Define the cross-system data alignment function, the expression is: A(v i ,v j )=tanh(W·[v i ;v j ]+b) Where W is the weight matrix and b is the bias vector; Construct the global alignment optimization objective, the expression is: Among them, K i,j represents the set of known aligned data pairs in systems i and j, and λ is the regularization parameter; Optimize L through the gradient descent method to obtain the optimal cross-system data alignment solution; Based on the alignment results, a unified device data representation is constructed to achieve the fusion of multi-source heterogeneous data.

4. The intelligent operation method of a commercial complex based on digital twins and IOT as claimed in claim 3, characterized in that: The automatic division of the inventory management task into multiple parallel subtasks comprises the following steps: Build a multi-dimensional task decomposition strategy based on equipment attributes and inventory characteristics; The inventory management problem is formalized as a multi-objective optimization problem, which considers the minimization of inventory cost and the maximization of service level at the same time; Genetic algorithm is used to generate the initial set of task decomposition solutions and the fitness function is designed. The expression is: F(x)=α·Cost(x)+β·ServiceLevel(x) Among them, x represents the task decomposition scheme, α and β are weight coefficients; Iteratively optimize the task decomposition scheme through crossover, mutation and selection operations until convergence or the maximum number of iterations is reached; For the generated optimal task decomposition solution, parallel computing technology is applied to execute multiple subtasks simultaneously; Set up a dynamic adjustment mechanism to adaptively adjust the task decomposition strategy based on real-time inventory status and execution results.

5. The intelligent operation method of a commercial complex based on digital twins and IOT as claimed in claim 4, characterized in that: The setting of the dynamic adjustment mechanism comprises the following steps: Define the state space S as the current inventory level and task execution progress, the action space A as possible task adjustment operations, and construct the reward function, which is expressed as: R(s,a,s′)=w1ΔCost+w2ΔServiceLevel+w3ΔEfficiency Among them, s and s' represent the status before and after adjustment, ΔCost, ΔServiceLevel and ΔEfficiency represent the changes in cost, service level and efficiency, respectively, and w1, w2 and w3 are weights; The deep Q learning algorithm is used to learn the optimal dynamic adjustment strategy through experience replay and target network technology. The loss function of the deep Q learning algorithm is defined as: L(θ)=E[(R+γmax a ′P(s′,a′;θ′)-P(s,a;θ)) 2 ] Among them, θ and θ' represent the parameters of the current network and the target network respectively, and γ is the discount factor; During the strategy execution process, a greedy strategy is used to balance exploration and utilization, and the strategy performance is evaluated regularly. If the performance degrades, the strategy retraining mechanism is triggered.

6. A smart operation system for a commercial complex based on digital twins and IOT, based on the smart operation method for a commercial complex based on digital twins and IOT according to any one of claims 1 to 5, characterized in that: include, The detection module is used to monitor the inventory status of equipment in real time using the IoT sensor network. When inventory changes are detected, the data is immediately transmitted to the cloud data center; An extraction module, used to perform intelligent classification procedures on the data received from the cloud and extract key feature information of the device; The construction module is used to build and dynamically update the intelligent inventory model of the equipment, match the extracted feature information in the model, and generate the corresponding inventory forecasting plan; A partitioning module is used to automatically divide the inventory management task into multiple parallel subtasks according to the generated inventory forecasting plan; The adjustment module is used to evaluate whether the conditions for automated inventory adjustment are met based on the matching results. If so, the adjustment is automatically performed. If not, the adjustment is performed after confirmation by the inventory manager. The forecasting module is used to apply time series forecasting algorithms, combine historical inventory data and usage frequency, predict future inventory demand, and provide data support for purchasing decisions.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, it implements the steps of any one of claims 1 to 5 of the method for intelligent operation of a commercial complex based on digital twins and IOT.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent operation method of a commercial complex based on digital twins and IOT as described in any one of claims 1 to 5 are implemented.

Images