Electronic component supply chain supply and demand management method

Abstract

The invention relates to the technical field of supply chain management, in particular to an electronic component supply chain supply and demand management method, which comprises the following steps: S1, abstractly tagging acquired electronic data; s2, selecting an LLM model, inputting the abstract tagging data in the S1 into the LLM model, analyzing the input data text by the LLM model, and outputting the result, the output result including further explanation of the supply and demand condition and prediction of the future supply and demand trend; s3, performing fine tuning on the LLM model by using a fine tuning module; s4, after system building is completed, vectorization representation is carried out on the labeled supply and demand data and the material data stored in the system, and rapid matching and query are carried out through a vector search technology; data at the supply end and the demand end are regularized through abstract labels, the consistency and efficiency of data processing are remarkably improved, and discrete and complex original data are converted into a unified label format by adopting an abstract label regularization technology.

Description

Technical Field

[0001] The invention relates to the technical field of supply chain management, in particular to a supply and demand management method for an electronic component supply chain. Background Art

[0002] Supply chain management of electronic components has always been a key link in the manufacturing and high-tech industries. With the continuous changes in market demand and the rapid iteration of technology, how to achieve efficient and accurate supply and demand matching has become a major challenge in supply chain management. Traditional supply chain management methods usually rely on experience and historical data for prediction and decision-making. For example, the existing management methods are generally MRP (Material Requirements Planning) system and ERP (Enterprise Resource Planning) system. The above management methods are overly dependent on past historical data. When the supply and demand relationship in the market fluctuates greatly, these historical data cannot determine the future supply and demand trend, and are usually based on batch processing and periodic updates. These methods also make it impossible to update the data. Therefore, the above two methods are often incapable of coping with rapid market changes and complex supply chain networks. Especially in the context of globalization, the complexity and uncertainty of the supply chain have further increased, resulting in frequent problems such as inventory backlogs, supply shortages, and delivery delays, which seriously affect the production efficiency and market competitiveness of enterprises.

[0003] In view of this, in order to overcome the above technical problems, the present invention proposes a supply and demand management method for an electronic component supply chain, which solves the above technical problems. Summary of the invention

[0004] In order to make up for the deficiencies of the prior art, the present invention proposes a supply and demand management method for an electronic component supply chain.

[0005] The technical solution adopted by the present invention to solve its technical problem is: A method for managing supply and demand of an electronic component supply chain, comprising: S1, abstract and label the acquired electronic data; S2, select the LLM model and input the abstract labeled data in S1 into the LLM model. The LLM model analyzes the input data text and outputs the results. The output results include further explanation of the supply and demand situation and prediction of future supply and demand trends; S3, fine-tune the LLM model using the fine-tuning module; S4, after the system is built, the labeled supply and demand data and material data stored in the system are vectorized and quickly matched and queried using vector search technology; S5, through Prompt engineering guidance system to dynamically adjust and optimize strategies to guide the model to generate specific output techniques and methods.

[0006] Furthermore, the electronic data abstract labeling includes performing labeling processing on the acquired data in a specific numerical range, specifically, determining the numerical range and the label.

[0007] Furthermore, the electronic data is obtained from the enterprise's internal system (such as ERP system) and external market data sources.

[0008] Furthermore, the electronic data is preprocessed, including deduplication, filling missing values ​​and outlier processing, and then feature extraction and label generation are performed.

[0009] Further, the labeled data is input into the LLM model through an API call or using an interface of an online platform.

[0010] Furthermore, the LLM model fine-tuning includes periodically updating the labels of the newly generated data and inputting them into the LLM model again to adjust the LLM model parameters, so as to observe whether the results output by the LLM model are consistent with the actual situation and the degree of consistency, so as to achieve the purpose of fine-tuning.

[0011] The beneficial effects of the present invention are as follows: The present invention realizes a significant improvement in the consistency and efficiency of data processing by regularizing the data on both ends of supply and demand through abstract labels. The abstract label regularization technology is used to convert discrete and complex raw data into a unified label format, which reduces errors and complexity in data processing and ensures the consistency and accuracy of data. Compared with the traditional method that relies on manual experience and historical data, this labeling process greatly simplifies the data management process and improves the processing efficiency. The present invention improves the accuracy and flexibility of supply and demand forecasts by introducing large models (such as large language models, LLM) and fine-tuning technology, and can better cope with the rapid changes in the market. For example, the large model is used for preliminary training, and personalized adjustments are made through fine-tuning technology to make the model more adaptable to the specific supply chain environment and demand changes. Compared with the traditional simple statistical model, this method greatly improves the complexity and adaptability of the forecast, and can provide more accurate forecasts in a rapidly changing market environment. The present invention adopts vector search and prompt prompt engineering technology to improve the real-time response capability and dynamic adjustment capability of the system, and uses vector search technology to quickly match and query data, and uses prompt prompt engineering to guide the system to make real-time adjustments and optimization strategies. Compared with traditional methods that rely on fixed rules and periodic adjustments, this technical approach enables the system to respond to changes in supply and demand in real time, adjust supply strategies in a timely manner, and ensure supply and demand balance. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention is further described below in conjunction with the accompanying drawings and implementation modes.

[0013] Figure 1 It is a schematic diagram of the framework structure of the working principle of the present invention. DETAILED DESCRIPTION

[0014] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0015] The electronic component supply chain supply and demand management method described in the present invention includes a data abstract label module and an LLM model tool, wherein the data is abstractly labeled and then input into the LMM model; It also includes tools for fine-tuning the LLM model, as well as vector search and prompt engineering modules; S1, abstract labeling of acquired electronic data: For example, the obtained data can be processed with specific numerical ranges and labels, specifically, the numerical ranges and labels can be determined, such as "demand fluctuations", such as "high demand (80-100%)", "medium demand (50-80%)", "low demand (0-50%)", and "supply fluctuations" can also be represented with specific numerical ranges; The data on both the supply and demand sides of electronic components can be obtained from the internal system of the enterprise (such as the ERP system) and the external market data source, and the obtained data can be pre-processed, including deduplication, filling missing values, outlier processing, etc., and finally feature extraction and label generation can be performed; among them, deduplication, filling missing values, outlier processing, etc. can be completed by plug-ins in the existing technology, so its working principle will not be described here in detail; Finally, the data is abstracted and labeled through the above operations. For example, for demand fluctuations, the corresponding label is determined according to the calculated demand growth rate; for example, if the demand growth rate is 85%, the label is "high demand"; For supply fluctuations, the supply status label can be determined by comparing the inventory level and delivery time with the set value range. The supply fluctuation can be "sufficient supply" or "tight supply"; By labeling specific numerical ranges, the fluctuations in supply and demand of electronic components can be quantified more accurately; S2, select the LLM model and input the abstract labeled data in S1 into the LLM model. The LLM model analyzes the input data text and outputs the results. The output results include further explanation of the supply and demand situation and prediction of future supply and demand trends: Input the labeled data into the Large Language Model (LLM). Select the appropriate LLM model and input the above sorted data into the LLM. You can input data through API calls or using the interface of the online platform. The LLM will analyze and generate output based on the input text. The output results include further explanation of the supply and demand situation and prediction of future trends, and analysis of the output results, such as combining the future forecast of supply and demand relationship with the actual business situation; S3, use the fine-tuning module to fine-tune the LLM model: The LLM model needs to be fine-tuned. As new data is continuously generated, it is necessary to update the labels of this data regularly and then input it into the LLM. This type of data is generated based on the market fluctuations unique to the enterprise, which can adjust the LLM model parameters, observe the consistency between the LLM output and the actual situation, and continuously adjust the data preparation and input methods to improve the accuracy and practicality of the LLM. S4, after the system is built, the labeled supply and demand data and material data stored in the system are vectorized and quickly matched and queried using vector search technology; S5, through prompt engineering guidance system to dynamically adjust and optimize strategies to guide the model to generate specific output techniques and methods.

[0016] In the description of the present invention, it should be noted that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate directions or positional relationships based on the attached Figure 1 The orientation or positional relationship shown is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, it cannot be understood as limiting the scope of protection of the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0017] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for managing supply and demand of an electronic component supply chain, characterized in that: include S1, abstract and label the acquired electronic data; S2, select the LLM model and input the abstract labeled data in S1 into the LLM model. The LLM model analyzes the input data text and outputs the results. The output results include further explanation of the supply and demand situation and prediction of future supply and demand trends; S3, fine-tune the LLM model using the fine-tuning module; S4, after the system is built, the labeled supply and demand data and material data stored in the system are vectorized and quickly matched and queried using vector search technology; S5, through Prompt engineering guidance system to dynamically adjust and optimize strategies to guide the model to generate specific output techniques and methods.

2. The electronic component supply chain supply and demand management method according to claim 1, characterized in that: The electronic data abstract labeling includes setting a specific numerical range for the acquired data and labeling the data, specifically, determining the numerical range and the label.

3. The electronic component supply chain supply and demand management method according to claim 2, characterized in that: The electronic data is obtained from the enterprise's internal system (such as ERP system) and external market data sources.

4. The electronic component supply chain supply and demand management method according to claim 3, characterized in that: The electronic data is preprocessed, including deduplication, filling missing values ​​and outlier processing, and then feature extraction and label generation are performed.

5. The electronic component supply chain supply and demand management method according to claim 1, characterized in that: The labeled data is input into the LLM model through an API call or using an interface of an online platform.

6. The electronic component supply chain supply and demand management method according to claim 1, characterized in that: The LLM model fine-tuning includes regularly updating the labels of the newly generated data and inputting them into the LLM model again to adjust the LLM model parameters, so as to observe whether the results output by the LLM model are consistent with the actual situation and the degree of consistency, so as to achieve the purpose of fine-tuning.

Images