Multi-merchant hardware equipment supply chain demand intelligent matching method

Abstract

The invention provides a multi-merchant hardware equipment supply chain demand intelligent matching method, and relates to the technical field of supply chain management, and the method comprises the steps: obtaining a demand order and supply capability data through a supply chain platform, and enabling the supply capability data to obtain production line test data in real time through an Internet of Things interface; performing preliminary screening based on equipment types and necessary authentication standards; calculating the comprehensive reliability of the optical fiber sensing system by adopting a comprehensive evaluation model fused with nonlinear transformation; calculating the multi-dimensional demand integrating degree of the supplier and the demander through multi-dimensional difference analysis and normalization processing; intelligent sorting is carried out based on weighted scores of configurable weight factors; implementing global productivity monitoring and conflict resolution optimization, and outputting an optimal matching pair; and continuously iteratively optimizing the evaluation model parameters according to the performance feedback data by adopting an incremental learning mechanism. According to the method, accurate matching and dynamic optimization under multi-target constraints are realized, and the supply chain resource configuration efficiency and the system autonomy capability are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of supply chain management, in particular to a multi-vendor hardware device supply chain demand intelligent matching method. BACKGROUND

[0002] In the development field of battery pack real-time damage detection and active protection system based on optical fiber sensing, the hardware device supply chain has high professionalism and complexity. The demand side usually needs to purchase precise detection devices of specific types and different performances, while the supply side provides diversified product lines and service capabilities. At present, the supply and demand matching of such hardware devices mainly depends on manual operation or basic information publishing platforms. The demand side publishes orders containing device types, performance requirements, budgets and delivery cycles, and the supply side displays its product information, but the subsequent screening and matching process is often inefficient and not accurate enough.

[0003] The existing matching method can only perform rough screening of device types and cannot perform fine evaluation of the comprehensive reliability of the supplier's devices and the degree of fit with the multi-dimensional requirements of the order. A sensor that can measure a specific parameter may have multiple suppliers, but its long-term stability in harsh working conditions, historical delivery reputation, and whether it can meet the specific cost and delivery requirements of the project are difficult to be quantitatively compared. This leads to a serious dependence on the personal experience of procurement personnel, which is highly subjective and makes it difficult to ensure the selection of the best supplier.

[0004] In addition, due to the lack of a global perspective on the real-time production capacity of suppliers, multiple demand orders may be matched to the same high-quality supplier, leading to excessive occupation of its production capacity and ultimately causing delivery delays and affecting the overall progress of downstream projects. At the same time, the existing static matching system lacks the ability to learn and optimize itself and cannot learn from historical transaction fulfillment data to continuously improve the accuracy and reliability of future matching, making it difficult to improve the overall efficiency of the supply chain. SUMMARY

[0005] To solve the technical problems of low efficiency and insufficient accuracy in the existing hardware device supply chain matching, lack of quantitative evaluation of the comprehensive reliability of suppliers and the fit of multi-dimensional requirements, inability to avoid capacity conflicts, and lack of self-learning and optimization capabilities of the system, the present application provides a multi-vendor hardware device supply chain demand intelligent matching method.

[0006] The technical solution provided by the present application is as follows:

[0007] The multi-vendor hardware device supply chain demand intelligent matching method provided by the present application comprises:

[0008] S1, supply chain demand data acquisition: through the supply chain management platform, device demand order data published by multiple hardware device demanders and device supply capacity data published by multiple hardware device suppliers are acquired; the device demand order data at least includes the type of required detection equipment, performance index requirement, budget range, and expected delivery period; the device supply capacity data at least includes the type of supplyable equipment, performance index parameter, offer, capacity state, and historical delivery credit score;

[0009] S2, preliminary screening of supply capacity and demand order: based on the matching of the equipment types in the device demand order data and the device supply capacity data, a candidate set of supply-demand equipment type matching is screened out;

[0010] S3, comprehensive reliability calculation of optical fiber sensing system: for each hardware device provided by the supplier screened in S2, based on its performance index parameter and historical delivery credit score, a comprehensive evaluation value representing the reliability and service quality of the hardware device provided by the supplier in subsequent actual application is calculated;

[0011] S4, multi-dimensional demand matching degree calculation: for each demand order and each supplier pair screened in S2, based on the performance index requirement, budget range, and expected delivery period of the demand order and the performance index parameter, offer, and capacity state of the supplier, a matching degree value representing the degree to which the supplier meets all aspects of the demand order requirements is calculated;

[0012] S5, intelligent matching and sorting: for each demand order, the pairing of the demand order and all potential suppliers is weighted and comprehensively scored according to the multi-dimensional demand matching degree and the comprehensive reliability of the optical fiber sensing system, and the sequence of recommended suppliers is generated according to the comprehensive score from high to low;

[0013] S6, matching result optimization and output: the recommended supplier sequence generated in S5 is subjected to conflict detection and optimization to ensure that the capacity of the same supplier is not over-allocated, and the optimal matching result is finally output to the supply chain management platform.

[0014] Further, the specific steps of calculating the comprehensive reliability of the optical fiber sensing system in step S3 are as follows:

[0015] S301, acquiring the performance index parameter of the hardware equipment provided by the supplier, the performance index parameter including the measurement accuracy, response time, and signal-to-noise ratio of the optical fiber sensor, and the control accuracy and response delay of the battery pack active protection system;

[0016] S302, acquiring the historical delivery credit score of the supplier;

[0017] S303, calculate the comprehensive reliability R of the optical fiber sensing system according to the following formula com :

[0018]

[0019] wherein R com represents the comprehensive reliability of the optical fiber sensing system; N represents the number of performance index types considered; P i represents the normalized value of the i-th performance index; S hist represents the historical delivery reputation score of the supplier; S min represents the minimum historical delivery reputation score allowed by the system; M represents the number of correction factors involved in the reliability calculation; C j represents the j-th correction factor, including but not limited to equipment average failure-free operation time, environmental adaptability level; α, β, γ are proportional coefficients obtained by training regression based on a large amount of supply chain data, and satisfy α+β+γ=1.

[0020] Further, the specific steps of calculating the multi-dimensional demand fit degree in step S4 are as follows:

[0021] S401, extract the performance index requirement vector D of the current demand order perf , budget requirement D budget , expected delivery period D cycle ;

[0022] S402, extract the performance index parameter vector S of the current supplier perf , offer S price , the expected delivery period corresponding to the capacity state S cycle ;

[0023] S403, calculate the multi-dimensional demand fit degree $F_{fit}$ according to the following formula:

[0024]

[0025] wherein F fit represents the multi-dimensional demand fit degree; represents the element-by-element difference operation between vectors; ||·|| represents the Euclidean norm of the vector; ||·||max represents the maximum value of the norm among all candidate pairs; |·| represents the absolute value; |·|max represents the maximum value of the absolute difference among all candidate pairs; ω1, ω2, ω3 are the preference coefficients of performance, price and delivery period respectively, which are determined by the demand side preset or dynamically analyzed by the system according to order data, and satisfy ω1+ω2+ω3=1.

[0026] Further, when obtaining the equipment supply capability data in step S1, current test data of the equipment on the production line of the supplier is also obtained in real time through the Internet of Things interface, and the current test data is taken as dynamic supplementary data of the performance index parameter.

[0027] Further, the preliminary screening in step S2 further includes: filtering the suppliers according to the necessary authentication standards specified in the demand order, and only keeping the suppliers with corresponding authentication qualifications in the candidate set.

[0028] Further, the weighted comprehensive score function in step S5 is:

[0029] Score = δ·R com +(1-δ)·F fit

[0030] Wherein, δ is a reliability weight factor set according to the preference of the demand side, and 0≤δ≤1.

[0031] Further, the value of the weight factor δ is explicitly specified by the demand side when issuing the order, and if not specified, the system defaults to δ = 0.5.

[0032] Further, the conflict detection and optimization in step S6 is specifically:

[0033] The system monitors the latest production capacity of each supplier in real time, and when it is detected that the predicted delivery period of a supplier is prolonged due to being matched by multiple high-priority orders and may affect order fulfillment, the system automatically starts a re-matching process for the affected demand order, or sends a negotiation prompt to the supply and demand sides.

[0034] Further, the method further includes the following steps after step S6:

[0035] S7, feedback learning and model updating: collecting actual fulfillment data of the orders that have completed matching and delivery, including actual performance of the equipment, on-time delivery rate, and after-sales service evaluation, and using the data to iteratively optimize the comprehensive reliability calculation model in step S3 and the multi-dimensional demand fit degree calculation model in step S4.

[0036] Further, the model updating in step S7 adopts an incremental learning method, and the model parameters are fine-tuned every pre-set period or after accumulating a certain amount of new feedback data, so as to adapt to the changes of the supply chain environment.

[0037] The technical scheme provided by the application has at least the following beneficial effects:

[0038] (1) In the present application, by constructing a comprehensive reliability and multi-dimensional demand matching degree quantitative model of the optical fiber sensing system, the overall and objective evaluation of the supply side and its equipment from long-term reliability to instant demand matching degree is realized, the subjectivity and one-sidedness of manual screening are overcome, and the precision and scientificity of the supply chain matching are significantly improved;

[0039] (2) In the present application, by introducing a conflict detection and optimization mechanism from a global perspective, the capacity allocation of the supply side is monitored and coordinated in real time, the delivery risk caused by multiple orders competing for the same high-quality capacity is effectively avoided, and the stability and performance reliability of the supply chain are ensured;

[0040] (3) In the present application, by establishing a feedback learning and model iterative updating mechanism based on actual performance data, the matching algorithm can continuously evolve and adapt to changes in the supply chain environment, thereby long-term improving the intelligence level and resource allocation efficiency of the entire supply chain system. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0042] Figure 1 A flowchart of a multi-merchant hardware device supply chain demand intelligent matching method provided in the embodiments of the present application. DETAILED DESCRIPTION

[0043] The technical solutions in the present application will be described below with reference to the drawings.

[0044] In the embodiments of the present application, the words such as "example", "for example" are used to represent as an example, illustration or explanation. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be either one of the two.

[0045] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent. "Of", "corresponding" and "corresponding" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent.

[0046] In the embodiments of the present application, sometimes the subscript such as W1 may be mistakenly written in the form of non-subscript such as W1, and when the difference is not emphasized, the meanings expressed are consistent.

[0047] To make the technical problems, technical solutions and advantages to be solved by the present application clearer, the following will be described in detail with reference to the drawings and specific embodiments.

[0048] Reference is made to the accompanying drawings Figure 1 , which shows a flowchart of a multi-merchant hardware device supply chain demand intelligent matching method provided by an embodiment of the present application.

[0049] The embodiment of the present application provides a multi-merchant hardware device supply chain demand intelligent matching method, and the processing flow can include the following steps:

[0050] S1, supply chain demand data acquisition: through a supply chain management platform, device demand order data published by a plurality of hardware device demanders is acquired, and device supply capacity data published by a plurality of hardware device suppliers is acquired; the device demand order data at least includes the type of the required detection device, the performance index requirement, the budget range, and the expected delivery cycle. The device supply capacity data at least includes the type of the supplyable device, the performance index parameter, the offer, the capacity state, and the historical delivery credit score.

[0051] It should be noted that in the specific implementation, the supply chain management platform provides a standardized data entry interface for the demander and the supplier. The demander submits the device demand order through a Web portal or an API interface, and when filling in the form, the application field of the required hardware device (such as a specific model of a fiber grating demodulator or a distributed optical fiber sensing host) needs to be clearly selected as battery pack monitoring, and the performance index thereof needs to be defined in detail, such as the number of sensing channels, the sampling rate, the dynamic range, the wavelength accuracy, the temperature and strain measurement accuracy, etc. The supplier maintains its product catalog, real-time capacity calendar, pricing strategy, and credit score automatically generated and updated by the platform algorithm according to historical transaction records in its background management system. The platform ensures the completeness and validity of the collected data through preset data verification rules, and all data are stored in the platform database in a structured manner to provide a data basis for subsequent intelligent matching.

[0052] In a possible implementation, when the device supply capacity data is acquired in step S1, the current test data of the device on the production line of the supplier is also acquired in real time through an Internet of Things interface, and the current test data is taken as dynamic supplementary data of the performance index parameter.

[0053] It should be noted that, in order to obtain the latest and most real device performance data, the system establishes an Internet of Things connection with the test station of the supplier's production line. Through an authorized API interface, the platform can obtain real-time test reports of devices just completed assembly and debugging on the assembly line at regular intervals (e.g., every minute) or triggered. These report data are automatically parsed and used to update the performance indicator parameters of the corresponding device model of the supplier. For example, the average value of the "signal-to-noise ratio" of the latest 100 devices is updated to the official published parameter value, or the measured distribution of the "measurement accuracy" is supplemented as a dynamic evaluation index to the supplier capability data. This makes the data on the platform no longer static, paper-based parameters, but dynamic data reflecting the current actual state of the production line, greatly improving the accuracy and reliability of matching.

[0054] Specifically, the Internet of Things interface adopts an industrial standard communication protocol based on MQTT or Modbus to ensure compatibility with production line test equipment of different manufacturers. The platform performs two-way authentication with the supplier system through a security certificate (such as an X.509 certificate) to establish an encrypted data transmission channel (such as TLS 1.2 or later) to ensure the security and integrity of real-time test data during transmission. The raw test data obtained (usually in JSON or XML format) is parsed and extracted by the platform's built-in parsing engine according to the predefined data schema, and finally converted into structured performance parameters and updated to the corresponding fields in the database.

[0055] S2, preliminary screening of supply capability and demand order: based on the device type in the device demand order data and the device supply capability data, the candidate set of the supplier and the demander device type matching is screened out.

[0056] It should be noted that this step is performed by the platform's background matching engine. The engine first parses the "device type" field in the demand order and the "suppliable device type" field in the supply capability data. In order to overcome the problem that different users may use different names for the same device, the platform maintains a standard dictionary of device types and a synonym mapping table to ensure that expressions such as "FBG demodulator", "fiber grating demodulator", etc. can be correctly normalized and recognized. The matching process first performs exact keyword matching, and for orders that cannot be matched exactly, the engine will use cosine similarity-based text similarity calculation for secondary expansion matching to cover a wider range of potential suppliers. Finally, all device type matching demanders and suppliers will be formed into a preliminary candidate matching set.

[0057] Specifically, the device type standard dictionary and the synonym mapping table are stored and maintained in the form of a graph database, where the nodes represent standard device types, and the edges represent "synonyms", "aliases", "subclasses", etc. relationships. The text similarity calculation based on cosine similarity, its text vector uses the term frequency-inverse document frequency (TF-IDF) model to generate, where the word library is based on the construction of the field technical documents and professional term library. The threshold of similarity calculation is set to 0.75, only when the similarity is higher than this threshold, the device type will be included in the candidate set, in this way, the recall rate and accuracy of matching are balanced.

[0058] In one possible implementation, the preliminary screening in step S2 also includes: filtering the suppliers according to the necessary certification standards specified in the demand order, and only keeping the suppliers with corresponding certification qualifications into the candidate set.

[0059] It should be noted that in the preliminary screening stage, in addition to device type matching, the system also enforces a certification qualification filtering. When the demand side issues an order, it can check one or more necessary certification standards in the form, such as "must pass ISO 9001 quality management system certification", "products must have CE certification" or "must have anti-explosion certification (ATEX / IECEx)", etc. When the matching engine performs screening, it will call the supplier qualification database to check whether the current supplier holds all the certifications required by the order. Only those suppliers with complete qualification files and within the valid period can pass this filtering condition and enter the subsequent candidate set. This mechanism ensures that the matched suppliers and their products meet the basic quality, safety or market access threshold.

[0060] S3, optical fiber sensing system comprehensive reliability calculation: for each supplier providing hardware devices after S2 screening, based on its performance index parameters and historical delivery credit score, a comprehensive evaluation value is calculated to represent the reliability and service quality of the hardware devices provided by the supplier in subsequent actual applications.

[0061] It should be noted that this step aims to quantitatively evaluate the overall reliability of the hardware devices and their services provided by the supplier. The platform extracts the detailed performance parameters of the devices provided by the candidate supplier from the database, which not only includes basic electrical characteristics, but also focuses on key performance in the battery pack monitoring scenario, such as sensor stability indicators in high temperature and humidity environments, accuracy retention rate under vibration conditions, etc. At the same time, the platform will obtain the historical delivery credit score of the supplier, which comprehensively considers factors such as the on-time delivery rate of its historical orders, the first-time machine failure rate, after-sales response speed, etc. The calculation process is not a simple weighted average, but a comprehensive evaluation model that integrates nonlinear transformation, aiming to amplify the contribution of high performance indicators and quality service records, so as to obtain a comprehensive reliability score that can significantly distinguish suppliers of different quality levels.

[0062] In one possible implementation, the specific steps of calculating the comprehensive reliability of the optical fiber sensing system in step S3 are as follows:

[0063] S301, obtain the performance indicator parameters of the hardware devices provided by the supplier, including the measurement accuracy, response time, signal-to-noise ratio of the optical fiber sensor, and the control accuracy, response delay of the battery pack active protection system;

[0064] S302, obtain the historical delivery credit score of the supplier;

[0065] S303, calculate the comprehensive reliability R of the optical fiber sensing system according to the following formula com :

[0066]

[0067] where R com represents the comprehensive reliability of the optical fiber sensing system; N represents the number of performance indicators considered; P i represents the normalized value of the i-th performance indicator; S hist represents the historical delivery credit score of the supplier; S min represents the minimum historical delivery credit score allowed by the system; M represents the number of correction factors involved in the reliability calculation; C j represents the j-th correction factor, including but not limited to the average trouble-free operation time of the device, the environmental adaptability level; α, β, γ are proportional coefficients obtained by training regression based on a large amount of supply chain data, and satisfy α+β+γ=1.

[0068] It should be noted that in the calculation of the comprehensive reliability R comWhen the system is initialized, it first reads the raw data of each performance indicator of the supplier's equipment from the standard fields of the database. For example, it reads the "measurement accuracy" (unit: % FS), "response time" (unit: ms), "signal-to-noise ratio" (unit: dB) from the "sensor" category, and "control accuracy", "response delay" from the "protection system" category. These raw data of different dimensions are first normalized to map them to the interval [0, 1] for comprehensive comparison. The normalized value of the i-th performance indicator is denoted as P i . At the same time, the system obtains its historical delivery reputation score S hist from the supplier management module, which is a value predefined in the range [0, 10] calculated by the platform according to historical behavior data. When calculating, the system calls the pre-installed algorithm module, and substitutes the normalized performance indicator array, historical reputation score, and fixed parameters (such as S min , a, b, g) read from the system configuration library into the formula for calculation. The correction factor C j in the formula, such as "mean time between failures (MTBF)", is automatically queried from the product archive library of the equipment to obtain its specific value.

[0069] It should be noted that the proportionality coefficients a, b, g are obtained by training using historical supply chain data (including a large number of completed equipment orders and their subsequent performance, delivery reputation, etc. data) using multiple linear regression or gradient descent algorithm. The goal of the training is to maximize the correlation between the calculation result of R com and the final actual satisfaction of the order (quantified by post-feedback data). The system loads a set of default regression coefficients at initialization, and continuously optimizes them in the feedback learning process of step S7. The "environmental adaptability level" in the correction factor C j is a normalized level score obtained by mapping according to a pre-defined query table based on the operating temperature range, humidity range, protection level (IP code) and other parameters provided by the equipment.

[0070] S4, Multi-dimensional demand fit degree calculation: for each pairing of the demand order and the supplier after S2 screening, based on the performance indicator requirements, budget range, expected delivery period of the demand order and the performance indicator parameters, offer, capacity status of the supplier, a fit degree value is calculated to represent the supplier's satisfaction of all aspects of the demand order.

[0071] It should be noted that this step is used to finely quantify the overall fit degree of a specific supplier to meet a specific demand order. The matching engine will process demand orders and suppliers in the candidate set pair by pair. For each pair combination, the engine will separately perform gap analysis from the performance, cost, and time dimensions. In the performance dimension, each performance index requirement of the demand party is compared with the corresponding parameter of the supplier equipment; in the cost dimension, the budget is compared with the offer; in the time dimension, the expected delivery period is compared with the predicted delivery period calculated by the supplier based on the current capacity state. The calculation process needs to normalize indicators of different dimensions and ranges, and comprehensively consider the demand party's possible preferences for different dimensions (such as some projects are extremely sensitive to performance, and some strictly control cost), and finally synthesize a fit degree score between 0 and 1, the higher the score, the better the matching degree.

[0072] In one possible implementation, the specific steps of calculating the multi-dimensional demand fit degree in step S4 are as follows:

[0073] S401, extract the performance index requirement vector D of the current demand order perf , budget requirement D budget , expected delivery period D cycle ;

[0074] S402, extract the performance index parameter vector S of the current supplier perf , offer S price , predicted delivery period S corresponding to the capacity state cycle ;

[0075] S403, calculate the multi-dimensional demand fit degree $F_{fit}$ according to the following formula:

[0076]

[0077] Where, F fit represents the multi-dimensional demand fit degree; represents the element-by-element difference operation between vectors; ||·|| represents the Euclidean norm of the vector; ||·||max represents the maximum value of the norm among all candidate pairs; |·| represents the absolute value; |·|max represents the maximum absolute difference value among all candidate pairs; ω1, ω2, ω3 are the preference coefficients of the performance, price, and delivery period dimensions, respectively, which are determined by the demand party preset or dynamically analyzed by the system according to the order data, and satisfy ω1+ω2+ω3=1.

[0078] It should be noted that the multi-dimensional demand fit degree F fitBefore the matching, the system first performs data vectorization. It combines the performance requirements in a demand order (e.g. measurement accuracy needs to be better than 0.1%, response time needs to be shorter than 5ms, etc.) into a demand performance vector D perf Similarly, it combines the performance parameters of the supplier devices into a supply performance vector S perf During the calculation, the system first calculates the absolute difference between the two vectors element by element, obtaining a difference vector. Then, it calculates the Euclidean norm of the difference vector to quantify the overall gap in performance between the two parties. In order to normalize, the system dynamically finds the maximum value of the norm within the current entire candidate matching set For the two scalar dimensions of price and delivery cycle, the absolute difference is also calculated, and the global maximum absolute difference is found. The default values of the preference coefficients ω1, ω2, ω3 are preset by the system administrator in the background strategy, for example, set to (0.5, 0.3, 0.2) to indicate that performance is more important. The demand party can adjust the distribution of the three coefficients through a sliding bar interface when submitting the order to meet its individual preferences.

[0079] Specifically, the element-by-element difference operation is For the performance index vector, the percentage of the absolute difference of the corresponding elements is defined as where k represents the kth dimension of the vector. For the dynamic finding of the norm |·|max, it is the global maximum value calculated in real time within the current candidate matching set filtered out in step S2, ensuring the fairness and context relevance of each matching calculation. If the preference coefficients ω1, ω2, ω3 are determined dynamically by the system, the logic is based on the historical order data of the demand party, identifying the preference pattern of the demand party (such as "performance sensitive type", "cost sensitive type" or "delivery urgent type") through cluster analysis, and automatically assigning the corresponding coefficient value.

[0080] S5, Intelligent matching and sorting: For each demand order, the pairing of all potential suppliers is weighted and scored according to the multi-dimensional demand fit degree and the comprehensive reliability of the fiber sensing system, and sorted from high to low according to the comprehensive score to generate a recommended supplier sequence.

[0081] It is to be noted that in this step, for each demand order to be processed, the matching engine calculates a final overall score for each supplier in its candidate list. This score is a linear weighted sum of the aforementioned two core metrics, “Fiber Sensing System Overall Reliability” and “Multi-dimensional Demand Fit”. The weight factor can be adjusted according to the preferences set by the demander when publishing the order (e.g. more emphasis on reliability or more emphasis on cost fit), or use the system default balanced weight if not specified. After the calculation, the engine sorts all potential suppliers for this demand order in descending order of the overall score, generating an ordered list of recommended suppliers, with the first one being the current optimal matching solution.

[0082] In one possible implementation, the weighted overall score function in step S5 is:

[0083] Score = δ · R com + (1 - δ) · F fit

[0084] where δ is the reliability weight factor set according to the demander’s preferences, and 0≤ δ≤ 1.

[0085] It is to be noted that the core of the intelligent matching and ranking is to calculate a single, comparable overall score (Score). This score function is designed as an adjustable weighted model. Among them, R com represents the long-term reliability and quality reputation of the supplier, F fit represents the instant fit of this transaction demand. The weight factor δ acts as a “slider” to determine which side is more biased in the final decision. This formula is embedded in the scoring module of the matching engine. For each (demand order, supplier) pair, the engine calls this formula, respectively inputs the calculated R com and F fit values, and the specified δ value, to quickly calculate its final score.

[0086] In one possible implementation, the value of the weight factor δ is explicitly specified by the demander when publishing the order, or the system defaults to δ = 0.5 if not specified.

[0087] It is noted that the assignment of the weight factor δ provides flexibility. On the order creation page, a special "matching strategy" option area is designed. Here, a clear slider is provided, with one end marked "most value reliability" and the other end marked "most value fit degree". The user can intuitively set the value of δ by dragging the slider (which can be mapped by the system to a value between 0 and 1). A text explanation is provided next to it to explain its meaning. If the demand side does not perform any operation and directly submits the order, the δ field in the order data will be assigned a system default value, which is usually set to 0.5, representing a balance between reliability and fit degree, ensuring that the matching process can produce a reasonable default result in any case.

[0088] S6, matching result optimization and output: the recommended supplier sequence generated in S5 is subjected to conflict detection and optimization to ensure that the capacity of the same supplier is not over-allocated, and finally the optimal matching result is output to the supply chain management platform.

[0089] It is noted that after the initial sorting list is generated, a round of global optimization is needed to avoid capacity conflicts. The platform system monitors the capacity reservation status of all suppliers in real time. When it detects that a high-quality supplier is simultaneously listed as the first preferred object by multiple high-priority orders, the system will predict whether its capacity is sufficient to meet all orders at the same time. If it is predicted that overload will occur and cause delivery delay, the system will automatically coordinate according to rules such as comprehensive score, order submission time, etc. It may select a sub-optimal but capacity-sufficient supplier for some orders, so as to ensure the stability and fulfillment reliability of the entire supply chain. Finally, the optimized matching result will be output to the demand side in the form of a list through the platform interface or message notification, and detailed comparison information will be provided for decision-making.

[0090] In one possible implementation, the conflict detection and optimization in step S6 are specifically:

[0091] The system monitors the latest capacity status of each supplier in real time. When it detects that a supplier's expected delivery period is extended due to being matched by multiple high-priority orders, which may affect order fulfillment, the system automatically starts a re-matching process for the affected demand orders, or sends a negotiation prompt to the supply and demand sides.

[0092] It is to be noted that the conflict detection is a continuous and dynamic process. The platform maintains a global "capacity reservation status table". After the matching engine generates a preliminary recommendation list, the optimization module simulates the allocation of the demand order to its first candidate supplier, and tries to reserve the corresponding capacity block for the supplier's capacity calendar in the status table. If this reservation causes the overall capacity utilization of the supplier in a certain time period to exceed the preset safety threshold (such as 95%), the system will trigger a conflict warning. The optimization module will not immediately reject the match, but will try to calculate the comprehensive score of the second candidate supplier for the order and check whether the allocation conflicts. If multiple orders conflict and cannot be automatically resolved, the system will generate a prompt message and send it to the demand side and the supply side respectively, informing them of the potential delivery risk and suggesting that they negotiate through the built-in communication tool of the platform.

[0093] Specifically, the "capacity reservation status table" is a time series-based database that records the reserved capacity and remaining capacity of each supplier in each time unit (such as day or week) in the future. The "safety threshold" is a configurable parameter, usually set by the system administrator according to industry experience, and the default is 85%. When simulating allocation, the optimization module uses a greedy algorithm or a constraint programming algorithm to try to find the second-best supplier with the most abundant remaining capacity for the orders affected by the conflict, while meeting the delivery period, so as to balance the global capacity load.

[0094] In one possible implementation, the method further comprises, after step S6:

[0095] S7, feedback learning and model updating: collecting actual performance data of completed matched and delivered orders, including actual performance of equipment, on-time delivery rate, after-sales service evaluation, using these data to iteratively optimize the comprehensive reliability calculation model in step S3 and the multi-dimensional demand fit calculation model in step S4.

[0096] Specifically, the method contains an important closed-loop learning mechanism. After the order is completed and delivered and put into operation for a period of time, the platform will collect feedback data through automated questionnaires and data analysis. These data include: actual performance data after the equipment is put into operation (compared with the claimed parameters), whether the delivery date is delayed, response time and solution efficiency of after-sales service requests, etc. These real performance data are stored in a special model training database. Periodically, the data analysis module of the platform uses these new feedback data to retrain the comprehensive reliability calculation model in step S3 and the multi-dimensional demand fit calculation model in step S4, so as to improve the accuracy of the two models. com and F fitThe mathematical model (for example, the coefficients a, b, g and w1, w2, w3 in the formula) is retrained and optimized, so that the model can evolve over time and more accurately predict the actual performance of the supplier and the degree of satisfaction of the demand.

[0097] In a possible implementation, the model updating in step S7 adopts an incremental learning manner, and the model parameters are fine-tuned every preset period or after a certain number of new feedback data are accumulated, so as to adapt to the changes in the supply chain environment.

[0098] It should be noted that the updating strategy of the model adopts an incremental learning manner instead of time-consuming full-retraining. Two triggering conditions are set: one is a time period (for example, a low peak period at the end of each week), and the other is a data amount threshold (for example, every time 100 new valid feedback records are accumulated). Once any condition is met, the incremental learning algorithm is started. Instead of discarding the old data, the new feedback data are taken as an incremental batch to fine-tune the existing model parameters. This way has high calculation efficiency, can quickly integrate the latest market changes and supplier performance into the model, enables the system to have self-adaptive capability, and avoids the model from producing sharp fluctuations due to a single batch of data, thereby ensuring the stability of the matching strategy.

[0099] Specifically, the incremental learning adopts an online gradient descent (Online Gradient Descent) or a stochastic gradient descent (Stochastic Gradient Descent) algorithm. The fine-tuning of the model parameters (for example, a, b, g, w1, w2, w3) aims to minimize a loss function defined as the mean squared error (Mean Squared Error) between the predicted value (for example, the calculated R com ,F fit ) and the real feedback value (for example, the actual performance score, the delivery satisfaction degree). The learning rate (Learning Rate) is set to a small value (for example, 0.001) to ensure stable convergence of the training process and avoid excessive disturbance of the model caused by new data.

[0100] The technical scheme provided by the embodiment of the application has at least the following beneficial effects:

[0101] (1) In the application, by constructing the optical fiber sensing system comprehensive reliability and multi-dimensional demand fit quantification model, the overall and objective evaluation of the supplier and the equipment from the long-term reliability to the instant demand matching degree is realized, the subjectivity and one-sidedness of the artificial screening are overcome, and the precision and scientificity of the supply chain matching are significantly improved.

[0102] (2) In the present application, by introducing the conflict detection and optimization mechanism of the global perspective, the capacity allocation of the supplier is monitored and coordinated in real time, effectively avoiding the delivery risk caused by multiple orders competing for the same high-quality capacity, and ensuring the stability and reliability of the supply chain;

[0103] (3) In the present application, by establishing a feedback learning and model iterative updating mechanism based on actual performance data, the matching algorithm can continuously evolve and adapt to changes in the supply chain environment, thereby long-term improving the intelligence level and resource allocation efficiency of the entire supply chain system.

[0104] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0105] The following points need to be explained:

[0106] (1) The drawings of the embodiments of the present application only involve the structures involved in the embodiments of the present application, and other structures can be referred to the usual design.

[0107] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present application, the thickness of the layers or regions is exaggerated or reduced, that is, the drawings are not drawn according to the actual proportion. It can be understood that when elements such as layers, films, regions or substrates are referred to as being located "on" or "under" another element, the element can be "directly" located "on" or "under" another element or there can be an intermediate element.

[0108] (3) In the case of no conflict, the embodiments of the present application and the features in the embodiments can be combined with each other to obtain new embodiments.

[0109] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for intelligent matching of supply chain demand for hardware equipment from multiple merchants, characterized in that, include: S1. Supply Chain Demand Data Acquisition: Through the supply chain management platform, acquire equipment demand order data published by multiple hardware equipment demanders, and simultaneously acquire equipment supply capacity data published by multiple hardware equipment suppliers; the equipment demand order data includes at least the type of testing equipment required, performance index requirements, budget range, and expected delivery cycle; the equipment supply capacity data includes at least the type of equipment available, performance index parameters, price, production capacity status, and historical delivery reputation score; S2. Preliminary screening of supply capacity and demand orders: Based on the equipment type in the equipment demand order data and equipment supply capacity data, a candidate set of equipment types that match the equipment types of the supplier and the demander is screened out; S3. Comprehensive Reliability Calculation of Fiber Optic Sensing System: For each supplier's hardware equipment after the S2 screening, a comprehensive evaluation value is calculated based on its performance index parameters and historical delivery reputation score to characterize the reliability and service quality of the hardware equipment provided by the supplier in subsequent practical applications. S4. Multidimensional Demand Fit Calculation: For each demand order and each supplier matched after S2 screening, a fit value is calculated based on the performance index requirements, budget range, expected delivery cycle of the demand order and the performance index parameters, quotation, and capacity status of the supplier, representing the degree of fit of the supplier in meeting all aspects of the demand order. S5. Intelligent Matching and Sorting: For each demand order, the matching with all potential suppliers is weighted and comprehensively scored based on the multi-dimensional demand fit and the comprehensive reliability of the fiber optic sensing system, and sorted from high to low according to the comprehensive score to generate a recommended supplier sequence. S6. Matching Result Optimization and Output: Conflict detection and optimization are performed on the recommended supplier sequence generated in S5 to ensure that the capacity of the same supplier is not over-allocated. Finally, the optimal matching result pair is output to the supply chain management platform.

2. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 1, characterized in that, The specific steps for calculating the overall reliability of the fiber optic sensing system in step S3 are as follows: S301. Obtain the performance parameters of the hardware equipment provided by the supplier, including the measurement accuracy, response time, and signal-to-noise ratio of the fiber optic sensor, as well as the control accuracy and response delay of the battery pack active protection system. S302. Obtain the supplier's historical delivery reputation score; S303. Calculate the overall reliability R of the fiber optic sensing system according to the following formula. com : Among them, R com Indicates the overall reliability of the fiber optic sensing system; N represents the number of performance indicators considered; P i S represents the normalized value of the i-th performance metric; hist This indicates the supplier's historical delivery reputation score; S min This represents the minimum historical delivery reputation score allowed by the system; M represents the number of correction factors involved in the reliability calculation; C j Let represent the j-th correction factor, including but not limited to the mean time between failures (MTBF) of the equipment and the environmental adaptability level; α, β, γ are the proportional coefficients obtained by training regression based on a large amount of supply chain data, and satisfy α+β+γ=1.

3. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 1, characterized in that, The specific steps for calculating the multidimensional demand fit in step S4 are as follows: S401. Extract the performance index requirement vector D of the current demand order. perf Budget requirement D budget Expected delivery cycle D cycle ; S402. Extract the current supplier's performance index parameter vector S perf Quotation S price The estimated delivery cycle S corresponding to the production capacity status cycle ; S403. Calculate the multidimensional demand fit degree $F_{fit}$ according to the following formula: Among them, F fit Indicates the degree of fit for multi-dimensional needs; ω1 represents the element-wise difference operation between vectors; ||·|| represents the Euclidean norm of the vector; ||·||max represents the maximum value of this norm among all candidate pairs; |·| represents the absolute value; |·|max represents the maximum value of this absolute difference among all candidate pairs; ω1, ω2, and ω3 are the preference coefficients for performance, price, and delivery cycle, respectively, which are preset by the demand side or dynamically determined by the system based on order data, and satisfy ω1+ω2+ω3=1.

4. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 1, characterized in that, In step S1, when acquiring equipment supply capacity data, the current test data of the equipment on the supplier's production line is also acquired in real time through the Internet of Things interface, and the current test data is used as dynamic supplementary data for performance index parameters.

5. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 1, characterized in that, The preliminary screening in step S2 also includes filtering suppliers according to the required certification standards specified in the demand order, and retaining only suppliers with the corresponding certification qualifications into the candidate set.

6. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 1, characterized in that, The weighted comprehensive scoring function in step S5 is: Score=δ·R com +(1-δ)·F fit Where δ is a reliability weighting factor set according to the demander's preferences, and 0≤δ≤1.

7. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 6, characterized in that, The value of the weighting factor δ is explicitly specified by the demander when posting an order. If not specified, the system defaults to δ = 0.

5.

8. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 1, characterized in that, The conflict detection and optimization in step S6 specifically involves: The system monitors the latest production capacity status of each supplier in real time. When it detects that a supplier's expected delivery cycle is extended due to being matched with multiple high-priority orders, which may affect order fulfillment, the system automatically initiates a re-matching process for the affected demand orders or sends a negotiation prompt to both the supplier and the demander.

9. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 1, characterized in that, The method further includes the following after step S6: S7. Feedback Learning and Model Update: Collect actual performance data of completed and delivered orders, including actual equipment performance, on-time delivery rate, and after-sales service evaluation. Use this data to iteratively optimize the comprehensive reliability calculation model in step S3 and the multi-dimensional demand fit calculation model in step S4.

10. The intelligent matching method for supply chain demand of multi-merchant hardware equipment according to claim 9, characterized in that, The model update in step S7 adopts an incremental learning approach. Every preset period or after accumulating a certain amount of new feedback data, the model parameters are fine-tuned to adapt to changes in the supply chain environment.

Images