Multi-agent material neat management system and method based on large language model

Abstract

The invention relates to a multi-agent material matching management system and method based on a large language model, and the system comprises a decomposition assistant agent which is used for carrying out the intelligent decomposition and recombination of material information in a supplier dimension; the draft assistant agent is used for generating communication texts oriented to different suppliers according to the intelligent decomposition and recombination result; the group sending assistant agent is used for sending the communication text to a corresponding supplier through a communication channel and tracking a feedback state of the supplier; and the tracking assistant agent is used for collecting and processing the feedback information according to the feedback information of the supplier. According to the invention, intelligentization and automation of material neat management are realized, the management efficiency is obviously improved, and the enterprise operation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of material set management in manufacturing industry, and in particular to a multi-agent material set management system and method based on a large language model. Background Art

[0002] The manufacturing industry is experiencing a critical period of transformation from traditional production mode to intelligent manufacturing. In this context, the importance of material set management as the core link of manufacturing production is becoming increasingly prominent. The efficiency of material set management directly affects the company's production plan execution, cost control and market responsiveness.

[0003] At present, large-scale manufacturing companies are facing multiple challenges in material set management. First, the product structure of modern manufacturing is becoming increasingly complex, and a single product often involves hundreds or even thousands of parts, which makes the collection and aggregation of material set information extremely arduous. Secondly, in the context of a global supply chain, companies need to maintain close communication with many suppliers at the same time and track the status of material supply in a timely manner, which not only requires a lot of manpower investment, but is also prone to problems such as information synchronization lag and low communication efficiency.

[0004] The traditional material set management method mainly relies on manual operation and basic information systems. Enterprises usually ensure the normal execution of production plans by setting up sufficient redundant capacity (including labor, equipment, production lines, etc.), flexibly adjusting overtime arrangements, and frequently communicating online and offline. This management method has the following obvious shortcomings:

[0005] First, manual operations account for too high a proportion. The collection, sorting, and analysis of material set information are mainly done manually, which is not only inefficient but also prone to human errors. Especially in scenarios with huge amounts of information and high timeliness requirements, the limitations of manual operations are more prominent.

[0006] Second, information synchronization is difficult. In the process of material set management, it is necessary to coordinate multiple departments within the enterprise and external suppliers. Information synchronization in this multi-party collaboration model is often delayed, affecting the timeliness and accuracy of decision-making.

[0007] Third, the ability to handle exceptions is limited. Traditional information systems are mainly based on predefined rules and processes. It is often difficult to respond quickly and flexibly to various abnormal situations that occur during the production process, and a lot of manual intervention is required.

[0008] Fourth, costs remain high. In order to cope with the uncertainty in material set management, companies have to maintain a high level of inventory and redundant production capacity, which directly leads to rising operating costs.

[0009] With the rapid development of artificial intelligence technology, especially large language model technology, a new technical path has been provided to solve the above problems. The large language model has a strong ability to understand and generate natural language. Combined with intelligent agent technology, it can realize the intelligence and automation of the material set management process, thereby improving management efficiency and reducing operating costs. This transformation can not only reduce manual operations and improve information processing efficiency, but also enhance exception handling capabilities through intelligent means, and achieve a qualitative improvement in material set management. Therefore, how to research and develop an intelligent material set management system based on a large language model has become an urgent problem to be solved. Summary of the invention

[0010] In order to solve the technical problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a multi-agent material set management system and method based on a large language model, which can realize the intelligent and automated material set management, and has the characteristics of timely information synchronization, high processing efficiency and low operating cost.

[0011] To achieve the above-mentioned object of the invention, the present invention provides a multi-agent material set management system based on a large language model, comprising:

[0012] Decomposition assistant agent, used to intelligently decompose and reorganize material information in the supplier dimension;

[0013] Drafting assistant agent, which is used to generate communication texts for different suppliers based on intelligent decomposition and reorganization results;

[0014] A group messaging assistant agent, used to send the communication text to the corresponding supplier through a communication channel and track the feedback status of the supplier;

[0015] The tracking assistant agent is used to collect and process the feedback information according to the feedback information of the supplier.

[0016] According to a technical solution of the present invention, the intelligent decomposition and reorganization of material information at the supplier dimension specifically includes:

[0017] Obtain the incoming material information and decompose the material information at the supplier dimension according to the downstream task requirements;

[0018] The decomposed material information is standardized to obtain standard material information particles, wherein the standard material information particles include material demand information, supplier information, material importance, and supply urgency;

[0019] Reorganizing the standard material information particles according to the supplier information of the standard material information particles;

[0020] According to the material importance and supply urgency of the standard material information particles, a priority label is assigned to the reorganized standard material information particles;

[0021] Output the intelligent decomposition and reorganization results.

[0022] According to a technical solution of the present invention, the drafting assistant agent generates communication texts for different suppliers, specifically including:

[0023] Build preset communication templates based on the type of material requirement scenario;

[0024] According to the supplier information, material requirement information and priority labels in the intelligent decomposition and reorganization results, the text content and tone of the communication template are dynamically adjusted to generate the communication text.

[0025] According to a technical solution of the present invention, the group messaging assistant agent tracks the feedback status of the supplier, specifically including:

[0026] Send the communication text to the corresponding supplier and mark the supplier's status as message sent;

[0027] According to the feedback information received from the supplier whose status is "message sent", the status of the supplier whose status is "message sent" is marked as message feedback, and the feedback information is sent to the tracking assistant agent.

[0028] According to a technical solution of the present invention, the tracking assistant agent collects and processes the feedback information, including: automatically collecting the feedback information of the supplier, automatically replying to the feedback information, and identifying and processing abnormalities in the feedback information of the supplier;

[0029] The abnormality identification and processing of the feedback information of the supplier specifically includes:

[0030] According to the abnormal corpus words in the feedback abnormal corpus, abnormality identification is performed on the feedback information of the supplier;

[0031] If there are abnormal corpus words, the supplier's feedback information will be copied to the corresponding person in charge according to the person in charge label of the abnormal corpus word, so as to manually reply to the feedback information.

[0032] According to a technical solution of the present invention, the decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent are all configured with:

[0033] The main function module is used to perform semantic analysis on the input natural language and output the corresponding functional manuscript;

[0034] The corpus module is used to store the corpus required by the agent to implement the main function;

[0035] The review module is used to optimize the functional manuscript output by the main functional module based on the ReAct automatic error correction mechanism.

[0036] According to a technical solution of the present invention, the optimization of the functional manuscript based on the ReAct automatic error correction mechanism specifically includes:

[0037] Observation phase: The agent observes and analyzes the functional manuscripts it outputs;

[0038] Reflection stage: Based on the preset quality standards, evaluate whether the output of the functional manuscript meets the requirements;

[0039] Correction stage: according to the output function manuscript cannot meet the requirements, the output function manuscript is iteratively optimized and a new function manuscript is regenerated;

[0040] Protection mechanism: Iterate optimization according to the set maximum number of error corrections to prevent the optimization of the intelligent agent from falling into an infinite loop.

[0041] According to a technical solution of the present invention, it also includes:

[0042] An agent backup module, used for backing up the decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent;

[0043] An agent status monitoring module, used for real-time monitoring of the operating status of the decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent;

[0044] The data backup module is used to regularly back up key data; the key data includes static data and dynamic data, the static data includes rule words and supplier information; the dynamic data includes material information and supplier status.

[0045] According to one solution of the present invention, a material set management method is implemented based on the above-mentioned multi-agent material set management system based on a large language model, comprising the following steps:

[0046] Step S1, the decomposition assistant agent intelligently decomposes and reorganizes the material information in the supplier dimension according to the received material information;

[0047] Step S2, the drafting assistant agent generates communication texts for different suppliers according to the results of the intelligent decomposition and reorganization;

[0048] Step S3, the group messaging assistant agent distributes the communication text to the corresponding supplier through the communication channel, and tracks the feedback status of the supplier;

[0049] Step S4: The tracking assistant agent collects and processes the feedback information according to the feedback information from the supplier.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] The intelligent material set management system of the present invention has achieved remarkable beneficial effects in practical applications by innovatively applying large language models and multi-agent collaborative technology, which are mainly reflected in the following aspects:

[0052] In terms of management efficiency, the present invention significantly improves the automation level of material set management. Through the multi-agent collaborative working mechanism, the system can automatically complete the decomposition of material information, document drafting, information sending and tracking, etc., greatly reducing the manual operation links. Especially when processing large quantities of material information, the advantages of the system are more obvious, and it can quickly complete the work that traditional manual methods take a long time to complete.

[0053] In terms of supply chain collaboration, the present invention improves the efficiency of information transmission through a unified communication mechanism. Based on a unified communication channel (such as WeChat public account, etc.), a standardized information transmission channel is established between enterprises and suppliers, which not only improves the collaborative efficiency of the supply chain, but also enhances the traceability of information transmission. The unified communication platform also enables suppliers to receive and respond to material demand information more conveniently.

[0054] In terms of exception handling, the present invention improves the efficiency of problem solving through an intelligent processing mechanism. The system can automatically handle common preset exceptions and issue early warnings in a timely manner for situations that require manual intervention. This rapid exception response mechanism enables enterprises to better control material supply risks and reduce the possibility of production interruptions.

[0055] In terms of data security, the present invention can adopt a localized deployment solution to ensure the security of enterprise data. By deploying core functions in the internal network of the enterprise, the system effectively prevents the risk of data leakage. At the same time, the stability of the system is also guaranteed, and the network connection is only used in necessary external communication links.

[0056] In terms of system reliability, the ReAct automatic error correction mechanism of the present invention ensures the accuracy of the processing results. Each agent has the ability to self-check and correct errors, which greatly reduces the workload of manual review and error correction. At the same time, the system's agent coordination mechanism ensures that the overall function can remain normal even when a single component has problems.

[0057] In general, the present invention provides manufacturing enterprises with an efficient and reliable material set management solution, which has obvious advantages in improving management efficiency, strengthening supply chain collaboration, etc., and provides strong support for the intelligent transformation of enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0059] Figure 1 Schematically showing the architecture diagram of a multi-agent material set management system based on a large language model provided in an embodiment of the present invention;

[0060] Figure 2 Schematically showing a structural diagram of a multi-agent material complete set management system based on a large language model provided in an embodiment of the present invention;

[0061] Figure 3 The flowchart of the multi-agent material set management method based on the large language model provided in an embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0062] The description of the embodiments of this specification should be combined with the corresponding drawings, which should be considered as part of the complete specification. In the drawings, the shape or thickness of the embodiments may be enlarged and indicated for simplification or convenience. Furthermore, the parts of each structure in the drawings will be described separately. It is worth noting that the elements not shown in the drawings or not described in words are in a form known to ordinary technicians in the relevant technical field.

[0063] The following description of the preferred embodiments may involve a combination of features, which may exist independently or in combination, and the present invention is not particularly limited to the preferred embodiments. The scope of the present invention is defined by the claims.

[0064] like Figure 1 As shown in the figure, the multi-agent material set management system based on a large language model of the present invention adopts a layered architecture design, which includes the application layer, the proxy layer, the RAG retrieval enhancement layer and the LLM core layer from top to bottom. This layered design not only ensures the modularity of the system, but also realizes the independent evolution and optimization of the functions of each layer.

[0065] As the top layer of the system, the application layer provides an intuitive user interface and supports concept verification and demonstration. At the same time, it provides the system with the ability to integrate with external applications through the API interface, ensuring the scalability of the system.

[0066] The agent layer is the core functional layer of the system, which includes multiple intelligent agent components working together. Each intelligent agent is an independent functional unit, which realizes information exchange and task collaboration through a unified communication protocol. The agent layer is responsible for processing natural language understanding and generation tasks, and supports the development of artificial intelligence applications based on agent technology.

[0067] The RAG search enhancement layer achieves deep integration between the system and the company's existing data. This layer uses core technologies such as word embedding, semantic retrieval, and vector database to establish an efficient knowledge retrieval mechanism. This enables the system to quickly retrieve relevant information from the company's knowledge base and combine it with preset prompt word templates to generate dynamic prompt words, thereby improving the system's question-answering capabilities and knowledge reserves in specific business scenarios.

[0068] The LLM core layer integrates a variety of open source and closed source large language models, including MOSS, Qwen2, Llama3.1, Deepseek, Baichuan2, YI, etc. Through model integration, the system can select the most suitable language model according to the characteristics of different tasks to achieve optimal performance.

[0069] like Figure 2 As shown, a multi-agent material set management system based on a large language model of the present invention includes:

[0070] Decomposition assistant agent, used to intelligently decompose and reorganize material information in the supplier dimension;

[0071] Drafting assistant agent, which is used to generate communication texts for different suppliers based on intelligent decomposition and reorganization results;

[0072] The group assistant agent is used to send communication texts to the corresponding suppliers through the communication channel and track the feedback status of the suppliers;

[0073] The tracking assistant agent is used to collect and process feedback information based on the feedback information from suppliers.

[0074] Among them, the decomposition assistant agent intelligently decomposes and reorganizes the material information in the supplier dimension, specifically including: obtaining the accessed material information, decomposing the material information in the supplier dimension according to the downstream task requirements; standardizing the decomposed material information to obtain standard material information particles, which contain material demand information, supplier information, material importance, and supply urgency; reorganizing the standard material information particles according to the supplier information of the standard material information particles; assigning priority labels to the reorganized standard material information particles according to the material importance and supply urgency of the standard material information particles; and outputting the intelligent decomposition and reorganization results. Through the above steps, the decomposition assistant agent can realize the decomposition and reorganization of material information (such as BOM information).

[0075] The drafting assistant agent receives the output of the decomposition assistant agent and generates communication texts for different suppliers, including: building preset communication templates according to the type of material requirement scenario; dynamically adjusting the text content and tone of the communication templates according to the supplier information, material requirement information and priority labels in the intelligent decomposition and reorganization results to generate communication texts.

[0076] The group assistant agent is responsible for batch sending and status tracking of information. The group assistant agent realizes unified communication with suppliers based on unified communication channels (such as WeChat public accounts, enterprise accounts, etc.), and adopts intelligent sending strategies to control the sending of supplier communication texts by controlling the sending frequency and time to avoid information bombardment. At the same time, the group assistant agent monitors the sending status of information in real time to ensure the reliability of communication, including: sending the communication text to the corresponding supplier, marking the supplier's status as sent; according to the feedback information received from the supplier whose status is sent, marking the supplier whose status is sent as message feedback, and sending the feedback information to the tracking assistant agent.

[0077] The tracking assistant agent collects and processes the supplier's feedback information through a unified communication channel, including: automatically collecting the supplier's feedback information, automatically replying to the feedback information, and identifying and processing the supplier's feedback information. Among them, the identification and processing of the supplier's feedback information includes: identifying the supplier's feedback information based on the abnormal corpus words in the feedback abnormality corpus; if there are abnormal corpus words, the supplier's feedback information is copied to the corresponding person in charge according to the person in charge label of the abnormal corpus word, so as to manually reply to the feedback information.

[0078] The present invention adopts a linear workflow mode to achieve orderly collaboration between agents. The decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent are all configured with:

[0079] The main function module is used to perform semantic analysis on the input natural language and output the corresponding functional manuscript;

[0080] The corpus module is used to store the corpus required by the agent to implement the main function;

[0081] The review module is used to optimize the functional manuscripts according to the output of the main functional module based on the ReAct automatic error correction mechanism.

[0082] The functional manuscript output by the decomposition assistant agent is the result of intelligent decomposition and reorganization; the functional manuscript output by the drafting assistant agent is the functional text; the functional manuscript output by the mass messaging assistant agent is feedback information; the functional manuscript output by the tracking assistant agent is an automatic reply to the feedback information, or feedback information copied to the corresponding person in charge.

[0083] The information transmission between agents is in the form of natural language, which ensures the flexibility and understandability of the system. The output of each agent is passed to the next agent only after verification. And the workflow of each agent supports conditional branching, which can select the appropriate processing path according to different situations.

[0084] Among them, the review module optimizes functional manuscripts based on the ReAct automatic error correction mechanism, including:

[0085] Observation phase: The agent observes and analyzes the functional manuscripts it outputs;

[0086] Reflection stage: Based on the preset quality standards, evaluate whether the output functional manuscript meets the requirements;

[0087] Correction stage: if the output functional manuscript fails to meet the requirements, the output functional manuscript is iteratively optimized and a new functional manuscript is regenerated;

[0088] Protection mechanism: Iterate optimization according to the set maximum number of error corrections to prevent the agent's optimization from falling into an infinite loop. The maximum number of error corrections can be set to 5 times. The ReAct automatic error correction mechanism can effectively improve the quality of the functional manuscripts output by each agent.

[0089] The multi-agent material complete set management system based on the large language model in the present invention also includes:

[0090] An agent backup module, used for backing up the decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent;

[0091] An agent status monitoring module is used to monitor the running status of the decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent in real time;

[0092] The data backup module is used to regularly back up key data; key data includes static data and dynamic data. Static data includes rule words and supplier information; dynamic data includes material information and supplier status.

[0093] The system provided by the present invention can be deployed online or offline. Local offline deployment has advantages in data security, network dependence and performance optimization: 1) Data security: core business data and models run within the enterprise to reduce the risk of data leakage; 2) Network dependence: the network is only used in necessary external communication links to improve system stability; 3) Performance optimization: local deployment reduces network latency and improves system response speed.

[0094] The present invention establishes an intelligent body redundancy mechanism through an intelligent body backup module, configures backup instances for key intelligent bodies, and ensures service continuity; constructs a status monitoring mechanism through an intelligent body status monitoring module, monitors the operating status of each component of the system in real time, and promptly discovers and handles abnormalities; constructs a data backup mechanism through a data backup module, regularly backs up key data, and prevents data loss.

[0095] In terms of data integration, for relatively stable static data, vector databases can be used for storage and management, including: data preprocessing: cleaning, standardization and vectorization of raw data; vector indexing: building efficient vector indexes to support fast similarity retrieval; on-demand updates: supporting administrators to manually trigger data updates to ensure the timeliness of data.

[0096] For real-time changing business data, the system provides the following two access methods: 1) Push mechanism: supports the enterprise's existing system to actively push data updates; 2) Interface call: provides a standardized interface to support real-time data query and acquisition.

[0097] like Figure 3 As shown, a material set management method provided by the present invention is implemented based on the above-mentioned multi-agent material set management system based on a large language model, and includes the following steps:

[0098] Step S1: The decomposition assistant agent intelligently decomposes and reorganizes the material information in the supplier dimension according to the received material information;

[0099] Step S2: The drafting assistant agent generates communication texts for different suppliers based on the results of intelligent decomposition and reorganization;

[0100] Step S3: The group messaging assistant agent sends the communication text to the corresponding supplier through the communication channel and tracks the feedback status of the supplier;

[0101] Step S4: The tracking assistant agent collects and processes the feedback information based on the supplier's feedback information.

[0102] It should be noted that the above is a preferred embodiment of the present invention. It should be pointed out that although the preferred embodiment of the present invention has been described, once the basic creative concept of the present invention is known, a number of improvements and modifications can be made by a person skilled in the art without departing from the principles of the present invention, and these improvements and modifications should also be regarded as the protection scope of the present invention. Therefore, the attached claims are intended to be interpreted as including the preferred embodiment and all changes and modifications that fall within the scope of the embodiments of the present invention.

Claims

1. A multi-agent material set management system based on a large language model, characterized in that: include: Decomposition assistant agent, used to intelligently decompose and reorganize material information in the supplier dimension; Drafting assistant agent, which is used to generate communication texts for different suppliers based on intelligent decomposition and reorganization results; A group messaging assistant agent, used to send the communication text to the corresponding supplier through a communication channel and track the feedback status of the supplier; The tracking assistant agent is used to collect and process the feedback information according to the feedback information of the supplier.

2. The multi-agent material set management system based on a large language model according to claim 1 is characterized in that: The intelligent decomposition and reorganization of material information at the supplier dimension specifically includes: Obtain the incoming material information and decompose the material information at the supplier dimension according to the downstream task requirements; The decomposed material information is standardized to obtain standard material information particles, wherein the standard material information particles include material demand information, supplier information, material importance, and supply urgency; Reorganizing the standard material information particles according to the supplier information of the standard material information particles; According to the material importance and supply urgency of the standard material information particles, a priority label is assigned to the reorganized standard material information particles; Output the intelligent decomposition and reorganization results.

3. The multi-agent material set management system based on a large language model according to claim 1 is characterized in that: The drafting assistant agent generates communication texts for different suppliers, including: Build preset communication templates based on the type of material requirement scenario; According to the supplier information, material requirement information and priority labels in the intelligent decomposition and reorganization results, the text content and tone of the communication template are dynamically adjusted to generate the communication text.

4. The multi-agent material set management system based on a large language model according to claim 1 is characterized in that: The group messaging assistant agent tracks the feedback status of the supplier, specifically including: Send the communication text to the corresponding supplier and mark the supplier's status as message sent; According to the feedback information received from the supplier whose status is "message sent", the status of the supplier whose status is "message sent" is marked as message feedback, and the feedback information is sent to the tracking assistant agent.

5. The multi-agent material set management system based on a large language model according to claim 1 is characterized in that: The tracking assistant agent collects and processes the feedback information, including: automatically collecting the feedback information of the supplier, automatically replying to the feedback information, and identifying and processing abnormalities in the feedback information of the supplier; The abnormality identification and processing of the feedback information of the supplier specifically includes: According to the abnormal corpus words in the feedback abnormal corpus, abnormality identification is performed on the feedback information of the supplier; If there are abnormal corpus words, the supplier's feedback information will be copied to the corresponding person in charge according to the person in charge label of the abnormal corpus word, so as to manually reply to the feedback information.

6. The multi-agent material set management system based on a large language model according to claim 1 is characterized in that: The decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent are all configured with: The main function module is used to perform semantic analysis on the input natural language and output the corresponding functional manuscript; The corpus module is used to store the corpus required by the agent to implement the main function; The review module is used to optimize the functional manuscript output by the main functional module based on the ReAct automatic error correction mechanism.

7. The multi-agent material set management system based on a large language model according to claim 6 is characterized in that: The optimization of the functional manuscript based on the ReAct automatic error correction mechanism specifically includes: Observation phase: The agent observes and analyzes the functional manuscripts it outputs; Reflection stage: Based on the preset quality standards, evaluate whether the output of the functional manuscript meets the requirements; Correction stage: according to the output function manuscript cannot meet the requirements, the output function manuscript is iteratively optimized and a new function manuscript is regenerated; Protection mechanism: Iterate optimization according to the set maximum number of error corrections to prevent the optimization of the intelligent agent from falling into an infinite loop.

8. The multi-agent material set management system based on a large language model according to claim 1 is characterized in that: Also includes: An agent backup module, used for backing up the decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent; An agent status monitoring module, used for real-time monitoring of the operating status of the decomposition assistant agent, the drafting assistant agent, the group sending assistant agent and the tracking assistant agent; The data backup module is used to regularly back up key data; the key data includes static data and dynamic data, the static data includes rule words and supplier information; the dynamic data includes material information and supplier status.

9. A material set management method, implemented based on the multi-agent material set management system based on a large language model as claimed in any one of claims 1 to 8, characterized in that: The following steps are involved: Step S1, the decomposition assistant agent intelligently decomposes and reorganizes the material information in the supplier dimension according to the received material information; Step S2, the drafting assistant agent generates communication texts for different suppliers according to the results of the intelligent decomposition and reorganization; Step S3, the group messaging assistant agent distributes the communication text to the corresponding supplier through the communication channel, and tracks the feedback status of the supplier; Step S4: The tracking assistant agent collects and processes the feedback information according to the feedback information from the supplier.

Images