Business resource dynamic optimization configuration method and system based on deep learning prediction

Abstract

The application belongs to the technical field of service resource allocation, and provides a service resource dynamic optimization configuration method and system based on deep learning prediction, acquires multi-source data, processes the multi-source data by standardization by using a predefined data template, extracts key associated fields of the multi-source data processed by standardization, constructs a data graph according to business logic, extracts business peak data associated with resource allocation in the data graph, extracts features of the business peak data, processes the extracted time sequence features, associated features and burst features by using a pre-trained deep learning model, predicts business data of a target period, and determines a resource allocation scheme of each type of resource according to a mapping rule library of a preset business scenario, resource type and quantitative standard relationship based on the predicted business data. The application can realize accurate prediction of business, and further realize accurate optimization configuration of service resources.

Description

Technical Field

[0001] This invention belongs to the field of business resource allocation technology, specifically relating to a method and system for dynamic optimization and allocation of business resources based on deep learning prediction. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Enterprise Resource Planning (ERP) software, as the core carrier of enterprise business processes, has accumulated massive amounts of business data (such as order data, production data, inventory data, financial data, etc.). The generation and flow of this data directly reflect the business load characteristics of the enterprise (such as a surge in order volume, centralized approval of production work orders, peak logistics documents, etc.). However, existing ERP systems have significant limitations in data value mining and decision support.

[0004] First, the ERP system adopts a modular architecture, with each module operating independently, resulting in scattered data storage and incompatibility with data formats and protocols of external systems, forming "data silos." This makes it difficult to form a comprehensive business view, resulting in the inability to accurately capture the correlation factors of business peaks. For example, sales order data cannot be correlated with production equipment capacity data and raw material inventory data in real time, leading to a lack of global data support for production planning.

[0005] On the other hand, ERP system forecasts rely on human experience or simple statistical methods (such as historical averages) and depend on the historical business volume of a single module. They do not consider the correlation of multi-source data and cannot adapt to the impact of sudden factors in industrial scenarios, resulting in large forecast errors and low forecast accuracy. Due to inaccurate forecasts, enterprises often face the problem of insufficient resources during business peaks and idle resources during troughs, leading to low operational efficiency and wasted resource costs. Summary of the Invention

[0006] To address the aforementioned problems, this invention proposes a method and system for dynamic optimization and allocation of business resources based on deep learning prediction. This invention is based on deep fusion of multi-source data, with deep learning prediction as the core and closed-loop resource allocation as the goal, to construct a full-link intelligent mechanism that can achieve accurate prediction of business needs, thereby enabling accurate optimization and allocation of business resources.

[0007] According to some embodiments, the present invention adopts the following technical solution:

[0008] A method for dynamically optimizing the allocation of business resources based on deep learning prediction includes the following steps:

[0009] Acquire multi-source data and perform standardization processing on the multi-source data using a predefined data template;

[0010] Extract key correlation fields from standardized multi-source data and construct a data graph according to business logic;

[0011] Extract peak business data related to resource allocation from the data map, and perform feature extraction on the peak business data. The extracted features include time-series features, correlation features, and burst features.

[0012] By using a pre-trained deep learning model, the extracted time-series features, correlation features, and burst features are processed to predict business data for the target time period.

[0013] Based on the predicted business data, and according to the pre-defined mapping rule library of business scenarios, resource types, and quantitative standard relationships, the allocation plan for various resources is determined.

[0014] As an alternative implementation method, the process of acquiring multi-source data includes: predefining an interface template library, using the interface template library to connect to external systems, and acquiring external related data;

[0015] Utilize a low-code configuration platform to receive registrations of new data sources via drag-and-drop operations.

[0016] Furthermore, the externally related data includes equipment operation data, warehousing operation efficiency data, customer demand and credit rating, real-time equipment status data, supply cycle and quality data, and market environment data.

[0017] As an alternative implementation method, the process of extracting key related fields from standardized multi-source data and constructing a data graph according to business logic includes: extracting key related fields, which include order number, material code and timestamp; and constructing a data graph according to business processes and business logic, which includes document characteristics, business scenarios and external influencing factors.

[0018] As an alternative implementation method, the process of extracting business peak data associated with resource allocation from the data map includes: extracting sales order peaks, raw material demand peaks, production load peaks, and inventory turnover peak data.

[0019] As an alternative implementation method, the process of extracting features from business peak data includes: extracting time-series features, including the periodic peak patterns of historical business data and the peak evolution in long-term trends;

[0020] Extract correlation features, including cross-business correlations that drive peak traffic, and the positive correlation between equipment failures and production order backlogs;

[0021] Extract sudden characteristics, including key factors that trigger non-periodic peaks.

[0022] As an alternative implementation method, the process of processing the extracted temporal features, correlation features, and burst features using a pre-trained deep learning model includes:

[0023] The pre-trained deep learning model includes a convolutional neural network and a long short-term memory network. The convolutional neural network is used for local feature extraction, and the long short-term memory network is used to deeply capture the temporal dependence of business peak data, accurately identify the periodic patterns of historical peaks, and introduce an attention mechanism to strengthen features that have a significant impact from peaks.

[0024] As a further implementation, the pre-trained deep learning model also incorporates enterprise resource planning (ERP) business rules to correct the output of the pre-trained deep learning model so that the output conforms to the ERP business rules.

[0025] As a further implementation, the parameters of the pre-trained deep learning model are dynamically adjusted based on the latest business data from the enterprise resource planning system after a set time period.

[0026] As an alternative implementation method, the process of determining the allocation plan for various resources based on predicted business data and according to a preset mapping rule library of business scenarios, resource types, and quantitative standard relationships includes: the preset mapping rule library of business scenarios, resource types, and quantitative standard relationships contains resources associated with various business peaks, as well as strategies for matching and calculating various resources; based on predicted business data and according to the mapping rule library, the quantity of each type of resource that needs to be matched is calculated to form an allocation plan; it is determined whether the quantity of each type of resource in the allocation plan exceeds a set threshold; if not, it is automatically sent to the corresponding business module; if so, a high-risk decision work order is generated for approval.

[0027] A system for dynamically optimizing and configuring business resources based on deep learning prediction, comprising:

[0028] The multi-source data processing module is configured to acquire multi-source data and perform standardization processing on the multi-source data using a predefined data template.

[0029] The data graph construction module is configured to extract key correlation fields from standardized multi-source data and construct data graphs according to business logic.

[0030] The peak feature extraction module is configured to extract business peak data related to resource allocation from the data map, and perform feature extraction on the business peak data. The extracted features include time-series features, correlation features, and burst features.

[0031] The business forecasting module is configured to use a pre-trained deep learning model to process the extracted time-series features, correlation features, and burst features to predict business data for the target time period.

[0032] The resource allocation module is configured to determine the allocation scheme for various resources based on the predicted business data and according to a pre-defined mapping rule library of business scenarios, resource types, and quantitative standard relationships.

[0033] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0034] This invention employs a combination of predefined interface templates and low-code configuration to enable direct access to internal and external ERP data, resolving issues of scattered data sources and heterogeneous formats. Furthermore, it constructs association rules based on core ERP business processes, automatically establishing cross-system data association graphs through key fields such as order numbers and material codes. This forms a fused data pool encompassing document characteristics, business scenarios, and external influences, ensuring deep data integration with business scenarios. Through format conversion, it guarantees the consistency and accuracy of the fused data, providing high-quality data support for subsequent processes.

[0035] This invention constructs a predictive model based on fused data, which combines time-series analysis capabilities with business adaptability. It accurately captures peak business patterns and mines three key features from multi-source data: time-series features, correlation features, and burst features, comprehensively covering peak driving factors. In prediction, it adopts a combination of convolutional neural networks, long short-term neural networks, attention mechanisms, and business rule correction. It utilizes the local feature extraction capabilities of convolutional neural networks and captures time-series dependencies through long short-term neural networks. The attention mechanism strengthens the weight of key features, and the business rule calibration adapts to the special patterns of industrial scenarios, significantly improving prediction accuracy.

[0036] This invention dynamically adjusts model parameters based on the latest business data at set intervals, enabling rapid response to emergencies such as urgent orders and equipment failures. It keeps the prediction error of core indicators within the set value, providing accurate time and magnitude basis for resource allocation.

[0037] This invention transforms prediction results into executable resource allocation instructions, forming a dynamic optimization closed loop of prediction-execution-feedback. Furthermore, based on the predicted business data, it establishes a rule base of business scenarios, resource types, and quantitative standards to achieve precise matching between resource demands and business pressures.

[0038] This invention achieves a progressive process of data fusion as the foundation, deep learning for accurate prediction, and closed-loop allocation to realize value, thereby upgrading enterprise resource planning from passive recording to proactive decision-making and improving enterprise resource utilization efficiency and operational intelligence.

[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0040] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0041] Figure 1 A flowchart illustrating one embodiment;

[0042] Figure 2 This is a schematic diagram of a deep learning model structure according to one embodiment. Detailed Implementation

[0043] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0044] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0045] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0046] Where there is no conflict, the embodiments and features described in this application may be combined with each other.

[0047] Example 1

[0048] As mentioned in the background section, the modular architecture of existing ERP systems leads to scattered data storage and incompatibility with external systems' data formats and protocols, which can easily create data silos.

[0049] Moreover, the data integration of existing ERP systems relies on customized development (such as hard-coded interfaces). When new data sources are added (such as introducing new supplier systems or adding production sensors) or the data source format changes, the interface needs to be redeveloped, which takes several weeks or even months. This cannot adapt to the dynamic needs of industrial enterprises such as production line adjustments and supply chain changes.

[0050] Existing ERP systems often rely heavily on manual experience or simple statistical methods (such as historical averages) for business forecasting. They depend on the historical business volume of a single module, failing to consider the correlation between multiple data sources (such as the impact of raw material price fluctuations on procurement volume and equipment failure rates on production schedules). Furthermore, they cannot adapt to the impact of unforeseen factors in industrial scenarios (such as urgent orders and equipment failures), resulting in large forecasting errors and low accuracy (e.g., peak order volume forecasting deviations exceeding 30%). Due to inaccurate forecasts, companies frequently face resource shortages during peak business periods (such as a shortage of approval personnel and logistics capacity) and idle resources during off-peak periods, leading to low operational efficiency (e.g., delayed order delivery) and wasted resources. Simultaneously, the decision-making process becomes disconnected from the forecast results. For example, after predicting a sales order peak, it fails to automatically trigger adjustments to production plans and logistics capacity allocation, resulting in resource shortages or idle resources.

[0051] Existing ERP systems rely heavily on built-in models from the product design for data processing. As businesses grow, the limitations of these built-in models become increasingly apparent. Furthermore, the sheer size and historical baggage of ERP systems make it difficult to quickly improve their data models and processing capabilities, hindering their ability to adapt to modern management systems and models.

[0052] To address the aforementioned issues of data silos, prediction lag, and inefficient resource allocation, this embodiment provides a method for dynamically optimizing the allocation of business resources based on deep learning prediction, such as... Figure 1 As shown, it includes the following steps:

[0053] Acquire multi-source data and perform standardization processing on the multi-source data using a predefined data template;

[0054] Extract key correlation fields from standardized multi-source data and construct a data graph according to business logic;

[0055] Extract peak business data related to resource allocation from the data map, and perform feature extraction on the peak business data. The extracted features include time-series features, correlation features, and burst features.

[0056] By using a pre-trained deep learning model, the extracted time-series features, correlation features, and burst features are processed to predict business data for the target time period.

[0057] Based on the predicted business data, and according to the pre-defined mapping rule library of business scenarios, resource types, and quantitative standard relationships, the allocation plan for various resources is determined.

[0058] In this embodiment, the process of acquiring multi-source data includes: taking full coverage, flexible adaptation and deep correlation as the core design principles, adopting a multi-dimensional and full-scenario access strategy at the data acquisition level to fully integrate two types of key data resources.

[0059] Firstly, the core business data within the ERP system covers all modules including procurement management (order quantity, delivery cycle, supplier information), production execution (work order process route, capacity demand, production progress), sales management (customer information, order demand, delivery deadline), inventory control (material code, inventory level, warehouse location distribution), and financial management (cost accounting, budget execution, accounts receivable and payable), fully covering basic attributes and core business fields such as business document number, timestamp, and status.

[0060] Secondly, external related data includes equipment operation data from the Manufacturing Execution System (such as overall equipment efficiency, failure rate, production load, etc.), warehousing operation efficiency from the Warehouse Management System (such as turnover rate, inbound and outbound timeliness, etc.), customer demand and credit rating from Customer Relationship Management, real-time equipment status collected by IoT sensors (such as temperature, energy consumption, operating parameters, etc.), supply cycle and quality data from supplier systems, and even external market environment data (such as industry promotional periods, raw material price fluctuations, changes in industry requirements, etc.).

[0061] In addition, in this embodiment, efficient integration is achieved through a three-layer adaptation mechanism:

[0062] First, there is a predefined interface template library, which develops standardized interfaces for mainstream external systems and supports one-click access configuration.

[0063] Second, there is a dynamic protocol adaptation engine. Users can select the corresponding data transmission protocol through a visual interface, and the corresponding format parser and timing data converter will be automatically loaded without manual coding.

[0064] Thirdly, the low-code configuration platform allows business personnel to complete the registration of new data sources (such as filling in access address, authentication information, update frequency, and other parameters) through drag-and-drop operations, compressing the traditional access cycle of several weeks to within 1 hour, and flexibly responding to the dynamic changes in data sources in scenarios such as production line adjustments and supplier changes for industrial enterprises.

[0065] Then, the data is fused and standardized.

[0066] Through a dynamic format conversion mechanism, data from different sources is standardized based on preset templates and custom rules (such as unit conversion rules and unified date formats), ensuring consistency in field naming, data types, and timestamp formats.

[0067] In this embodiment, relying on key related fields such as order number, material code, and timestamp, and following business logic such as "sales order → production work order → raw material procurement → supplier" and "production work order → equipment resources → energy consumption data", a multi-source data association graph is automatically constructed to form a fusion data pool containing "document characteristics - business scenario - external influencing factors", forming a visualized business data graph that supports tracing the upstream and downstream data of any business node.

[0068] Through the above solution, this embodiment can achieve "plug-and-play" access to multi-source data inside and outside the ERP system, solve the problems of scattered data sources and heterogeneous formats, and realize deep integration from isolated data to data across the entire business chain, providing complete, consistent, and related high-quality data support for subsequent dynamic prediction and intelligent decision-making.

[0069] In the data prediction process, pre-trained deep learning models are used to accurately predict peak business data or determine peak business data.

[0070] The deep learning prediction layer takes multi-source fusion data as its core input and uses a specially designed deep learning model and dynamic optimization mechanism to accurately predict peak data in the core business of the ERP system, providing enterprises with decision-making basis for responding to business fluctuations in advance.

[0071] Extract core business data closely related to resource allocation from the ERP system, including: peak sales order periods (such as peak order dates, volumes, and the proportion of orders from key customers in the next 7 / 30 days), peak raw material demand periods (such as peak procurement volumes and potential shortages within 15 days), peak production load periods (such as peak work order saturation rates for each production line within one week and critical points for equipment utilization), and peak inventory turnover periods (such as peak consumption periods for key materials and stockout warning windows). These data directly reflect the concentrated points of business pressure, and their forecasting results are the core basis for enterprises to allocate resources in advance.

[0072] The data fusion analysis identifies three key characteristics strongly correlated with business peaks: First, temporal characteristics: periodic peak patterns in historical business data (e.g., a surge in sales orders every Monday, or a concentrated burst of production orders at the end of a quarter), and the evolution of peaks in long-term trends (e.g., the year-on-year increase in peak season order levels); second, correlation characteristics: cross-business correlations driving peaks (e.g., a surge in purchase orders triggered by a sharp drop in raw material prices, or a chain reaction of sales-logistics order peaks caused by promotional activities), and a positive correlation between equipment failures and backlogs of production orders; and third, sudden characteristics: key factors triggering non-periodic peaks (e.g., a sharp increase in the proportion of urgent orders, concentrated replenishment of production orders due to supplier delays, or temporary additional orders from customers).

[0073] In addition, this embodiment also utilizes a deep learning model to optimize the accuracy of peak prediction.

[0074] This embodiment adopts a model architecture consisting of convolutional neural network + attention mechanism + long short-term memory network + business rule calibration.

[0075] Specifically, such as Figure 2 As shown, it includes an input layer, a convolutional neural network layer, a long short-term memory network layer, an attention layer, a fully connected layer, and an output layer arranged in sequence.

[0076] The input layer receives normalized and standardized multi-source data, which can be divided into datasets {X1, X2, ..., X...}. n}

[0077] The convolutional neural network layer utilizes the local feature extraction capability of the convolutional neural network (CNN), while the long short-term memory network layer deeply captures the temporal dependence of business data through several long short-term memory (LSTM) networks, accurately identifying the periodic patterns of historical peaks (such as the peak sales season each year).

[0078] Convolutional neural networks consist of convolutional layers and pooling layers, which can be achieved using existing technologies and will not be elaborated upon here. Long short-term memory networks can also be achieved using existing technologies and will not be elaborated upon here.

[0079] The attention layer introduces an attention mechanism to emphasize features that have a significant impact on peaks (such as the weight of promotional activities on peak sales orders, the impact of equipment maintenance plans on peak production loads, etc.). This layer can use the Softmax function as the activation function.

[0080] Next, after dense connections are made through a fully connected layer, the output is then output through the output layer.

[0081] A calibration layer can also be introduced, incorporating ERP business rules (such as "order volume increased by 30% month-on-month in the two weeks before the Spring Festival" and "the peak formed by major customers placing orders on the 5th of each month"), to correct the output of the deep learning model and adapt to regular sudden peaks in industrial scenarios.

[0082] To ensure adaptability to dynamic business, the deep learning model establishes a real-time update mechanism: In this embodiment, the parameters of the model (such as the weight of the attention mechanism and the model's periodic factor) are dynamically adjusted and optimized every 6 hours based on the latest business data from the ERP (such as newly generated emergency orders, equipment failure records, material warehousing information, etc.) to quickly respond to unexpected peaks caused by sudden factors (such as emergency order insertion, equipment shutdown).

[0083] Through this mechanism, the prediction error of peak core business data can be controlled within 10%. The prediction accuracy of key indicators such as peak sales orders and critical production load points is particularly outstanding, providing precise timing and magnitude basis for enterprises to adjust production plans in advance, reserve raw materials, and allocate logistics capacity.

[0084] Based on peak business data output by the deep learning prediction layer, dynamic adaptation of resources to business needs is achieved through rule-based matching, phased strategies, and closed-loop feedback mechanisms, thereby improving the resource utilization efficiency of the ERP system.

[0085] The resource demand matching mechanism focuses on the precise correspondence between business peaks and resource types: a mapping rule library of business scenarios, resource types, and quantitative standards is established based on prediction results.

[0086] For example, the mapping rule base shows the logistics capacity and approval manpower resources corresponding to peak sales orders (>500 orders / day), with a standard configuration of 2 approvers and 3 delivery vehicles required for 100 orders.

[0087] In the mapping rule base, peak production load (production line saturation rate > 90%) is associated with two types of resources: equipment and skilled workers. Each piece of equipment is matched with 3 skilled workers.

[0088] When the raw material demand gap exceeds a set value (e.g., gap > 500 units), both procurement and inventory resources are triggered to respond, and a purchase order is generated according to the formula of gap amount × 1.2 safety factor.

[0089] The mapping rule base supports customization according to ERP business processes, such as inventory allocation rules for cross-regional sales (e.g., when orders increase by 50%, 30% safety stock is allocated from the sufficient warehouse).

[0090] The generated dynamic allocation strategy can be executed in layers according to the time dimension, ensuring the timeliness and foresight of resource response. For example, the short-term (1-3 days) strategy targets emergency peaks and directly generates ERP executable instructions: such as pushing approval work orders to idle personnel through the application programming interface (API), calling on external logistics capacity to make up for the gap, and temporarily adjusting the production line load (scheduling saturated work orders to idle equipment).

[0091] Mid-term (e.g., 1-4 weeks) strategies focus on planned adjustments: advance equipment maintenance based on production peak forecasts (e.g., avoiding peak-hour breakdowns), optimize inventory layout based on sales trends (e.g., moving popular products to regional warehouses), and develop technician scheduling plans (e.g., increasing manpower by 20% on peak days), etc.

[0092] The allocation strategy is executed in two modes: automatic and approval. Standardized operations (such as routine procurement and inventory transfer) are implemented directly by the system, i.e., in automatic mode, and are automatically sent to the corresponding business modules for execution. High-risk decisions (such as emergency procurement exceeding the budget or resource allocation exceeding the set value) can be pushed to the ERP workbench for approval before execution.

[0093] In some embodiments, real-time data on the adjusted business (such as order delivery timeliness, equipment utilization rate, procurement arrival rate, etc.) can also be collected and compared with the expected targets.

[0094] For example, if 3 approvers are needed to verify 100 sales orders, the mapping rule base will be automatically updated. For example, if it is found that 4 delivery vehicles are needed for 500 orders / day, the logistics capacity standard will be adjusted; if the deviation between the predicted and actual raw material shortage exceeds 15%, the safety factor and the parameters of the deep learning model used for prediction will be adjusted in reverse.

[0095] Through this mechanism, the matching degree between resource allocation and business needs is continuously improved, which avoids resource shortages during peak hours and reduces idle waste during off-peak hours, realizing the intelligent and dynamic scheduling of ERP system resources.

[0096] Example 2

[0097] A system for dynamically optimizing and configuring business resources based on deep learning prediction, comprising:

[0098] The multi-source data processing module is configured to acquire multi-source data and perform standardization processing on the multi-source data using a predefined data template.

[0099] The data graph construction module is configured to extract key correlation fields from standardized multi-source data and construct data graphs according to business logic.

[0100] The peak feature extraction module is configured to extract business peak data related to resource allocation from the data map, and perform feature extraction on the business peak data. The extracted features include time-series features, correlation features, and burst features.

[0101] The business forecasting module is configured to use a pre-trained deep learning model to process the extracted time-series features, correlation features, and burst features to predict business data for the target time period.

[0102] The resource allocation module is configured to determine the allocation scheme for various resources based on the predicted business data and according to a pre-defined mapping rule library of business scenarios, resource types, and quantitative standard relationships.

[0103] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of one or more computer-usable storage media (including, but not limited to, disk storage, etc.) containing computer-usable program code. CD - ROM It takes the form of a computer program product implemented on (such as optical memory, etc.).

[0104] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0107] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for dynamically optimizing configuration of service resources based on deep learning prediction, characterized in that, The method comprises the following steps: acquiring multi-source data, and performing standardization processing on the multi-source data by using a predefined data template; the process of acquiring multi-source data comprises: predefining an interface template library, using the interface template library to interface with external systems, and acquiring external associated data; using a low-code configuration platform to receive registration of a new data source through a drag-and-drop operation; and using a dynamic format conversion mechanism to standardize data from different sources based on a preset template and a custom rule, so as to ensure consistency of field naming, data type, and timestamp format; extracting key associated fields of the multi-source data after standardization processing, and constructing a data graph according to business logic; the process of extracting key associated fields of the multi-source data after standardization processing and constructing a data graph according to business logic comprises: extracting key associated fields, wherein the key associated fields comprise order number, material code, and timestamp; constructing a data graph according to business process and business logic, wherein the data graph comprises document features, business scenarios, and external influencing factors; extracting business peak data associated with resource allocation in the data graph, and performing feature extraction on the business peak data, wherein the extracted features comprise time sequence features, associated features, and burst features; using a pre-trained deep learning model to process the extracted time sequence features, associated features, and burst features, and predicting business data in a target period; the pre-trained deep learning model further introduces enterprise resource planning business rules to correct the output of the pre-trained deep learning model, so that the output conforms to the enterprise resource planning business rules; parameters of the pre-trained deep learning model are dynamically adjusted based on the latest enterprise resource planning business data after a set period of time; determining an allocation scheme for each type of resource according to the predicted business data and a preset mapping rule library of business scenarios, resource types, and quantitative standard relationships; the process of performing feature extraction on the business peak data comprises: extracting time sequence features, including periodic peak rules of historical business data and evolution of peaks in long-term trends; extracting associated features, including cross-business correlation driving peaks and positive correlation between equipment failure and production order backlog; extracting burst features, including key factors triggering non-periodic peaks; the process of determining an allocation scheme for each type of resource according to the predicted business data and the preset mapping rule library of business scenarios, resource types, and quantitative standard relationships comprises: the preset mapping rule library of business scenarios, resource types, and quantitative standard relationships comprises various resources associated with business peaks and strategies for matching calculation of each type of resource; according to the predicted business data, the number of each type of resource that needs to be matched is calculated according to the mapping rule library, forming an allocation scheme, and determining whether the number of each type of resource in the allocation scheme exceeds a set threshold; if not, the allocation scheme is automatically issued to the corresponding business module; if yes, a high-risk decision work order is generated for approval. 2.The method of claim 1, wherein the method further comprises: determining a resource allocation scheme for each of the plurality of services based on the predicted resource demand of each of the plurality of services. The external associated data comprises equipment operation data, warehouse operation efficiency data, customer demand and credit rating, real-time equipment state data, supply cycle and quality data, and market environment data. 3.The method of claim 1, wherein the method further comprises: determining a resource allocation scheme for each of the plurality of services based on the predicted service resource demand and the predicted service resource supply; and configuring the resource allocation scheme for each of the plurality of services. The process of extracting the business peak data associated with resource allocation in the data graph includes extracting sales order peak, raw material demand peak, production load peak, and inventory turnover peak data.

4. The method of claim 1, wherein the method further comprises: determining a resource allocation for each of the plurality of services based on the predicted resource demand for each of the plurality of services. The process of processing the extracted time series features, correlation features, and burst features using the pre-trained deep learning model includes: The pre-trained deep learning model includes a convolutional neural network for local feature extraction and a long short-term memory network for deep capture of time series dependence of business peak data, accurate identification of periodic patterns of historical peaks, and introduction of an attention mechanism for feature enhancement that significantly affects the peak.

5. A system for dynamically optimizing configuration of business resources based on deep learning prediction, applying the method of claim 1, characterized in that, It includes: A multi-source data processing module configured to obtain multi-source data, and to standardize the multi-source data using a predefined data template; The process of obtaining multi-source data includes: a predefined interface template library, using the interface template library to interface with external systems to obtain external associated data; using a low-code configuration platform to receive registration of new data sources through drag-and-drop operations; and standardizing data from different sources based on pre-defined templates and custom rules to ensure consistency in field naming, data types, and timestamp formats; A data graph construction module configured to extract key associated fields of the standardized multi-source data and construct a data graph according to business logic; the process of extracting key associated fields of the standardized multi-source data and constructing a data graph according to business logic includes: extracting key associated fields including order number, material code, and timestamp, and constructing a data graph according to business processes and business logic, the data graph containing document features, business scenarios, and external influencing factors; A peak feature extraction module configured to extract business peak data associated with resource allocation in the data graph and perform feature extraction on the business peak data, the extracted features including time series features, correlation features, and burst features; A business prediction module configured to process the extracted time series features, correlation features, and burst features using a pre-trained deep learning model, and to predict business data in a target period; the pre-trained deep learning model also incorporates enterprise resource planning business rules to correct the output of the pre-trained deep learning model, so that the output conforms to the enterprise resource planning business rules; The parameters of the pre-trained deep learning model are dynamically adjusted based on the latest enterprise resource planning business data after a set period of time; A resource allocation module configured to determine resource allocation schemes for various types of resources based on predicted business data, according to a pre-set mapping rule library of business scenarios, resource types, and quantitative standard relationships.

Images