Method and system for artificial intelligence based supply chain optimization decisions

Abstract

The application relates to the technical field of supply chain management, and discloses a method and system for supply chain optimization decision based on artificial intelligence, which comprises a customized service module, a data analysis module, an intelligent evaluation module and an optimization decision module. The method and system for supply chain optimization decision based on artificial intelligence acquire the management data of suppliers and manufacturers and the demand data of automobile refitting part orders in real time through the customized service module, and classify and form data sets, the data analysis module calculates the supply coefficient and risk value of each supplier and the assembly precision and quality coefficient of each manufacturer, forms a supplier ranking database and a manufacturer ranking database, the intelligent evaluation module judges the complexity level of the automobile refitting part, further screens out line data groups meeting the automobile refitting part order demand, and calculates the customization duration; the intelligent configuration production line has high flexibility; the optimization decision module monitors the actual production situation, carries out corresponding optimization measures, and dynamically adjusts the product quality to be more stable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of supply chain management, in particular to a method and system for supply chain optimization decision based on artificial intelligence. BACKGROUND

[0002] The supply chain, as the core of enterprise operation, refers to a series of value-added activities from raw material procurement to final product delivery to consumers. Participants in the supply chain include suppliers, manufacturers, distributors, retailers and end consumers. The management process of the supply chain involves planning, procurement, production, transportation, warehousing and customer service, etc. Through optimizing the supply chain, enterprises can establish long-term cooperation with high-quality suppliers and manufacturers, and achieve lean production.

[0003] With the change of market environment and the upgrading of consumer demand, especially in the field of automobile modification parts, consumers' demand for automobile modification parts increasingly presents the characteristics of small batch, diversification and individualization, which poses new challenges to traditional supply chain, but also brings opportunities for transformation and upgrading of enterprises. By introducing digital technology, optimizing production mode, improving collaboration ability and strengthening customer interaction, enterprises can build an efficient, flexible and transparent supply chain system, thereby better serving consumers. In the future, with the further development of technology and the continuous growth of market demand, the optimization of automobile modification parts supply chain will become an important source of industry competitiveness.

[0004] Currently, traditional supply chain often focuses on mass production, lacking flexibility to respond to small batch, individualized customization orders. This rigid production mode is prone to cause modification parts quality not meeting consumer expectations, delayed delivery time, etc., making it difficult to meet consumers' time requirements. In addition, in the actual operation process of the supply chain, information and demand of different nodes may change rapidly, and traditional supply chain lacks real-time data sharing and dynamic adjustment capability, resulting in resource allocation and production planning unable to update in time, affecting the decision-making efficiency of the overall supply chain and consumer satisfaction. SUMMARY

[0005] (I) Technical problems solved

[0006] In view of the shortcomings of the prior art, the present application provides a method and system for supply chain optimization decision based on artificial intelligence, which has the advantages of high flexibility of intelligent configuration production line, stable dynamic adjustment of product quality, etc., and solves the problems of traditional supply chain that is difficult to meet the needs of small amount of individual customization and lacks dynamic adjustment capability.

[0007] (II) Technical solutions

[0008] To achieve the above-mentioned purpose, the present application provides the following technical solutions: a method for supply chain optimization decision based on artificial intelligence, comprising the following steps:

[0009] Step one: Real-time access to all suppliers and manufacturers' management data and all automotive modification parts order demand data through network connection supply chain management system, and classify them into supply data set, manufacturing data set and order data set;

[0010] Step two: According to the supply data set, analyze the delivery rate JV, the change trend of the delivery rate JV and the delivery risk of each supplier, generate the corresponding supply coefficient Gyxs and risk value Fxz, and form the supplier ranking database GK;

[0011] Step three: According to the manufacturing data set, analyze the assembly accuracy Zujd and quality coefficient Zlxs of each manufacturer, and form the manufacturer ranking database ZK;

[0012] Step four: According to the supply data set, manufacturing data set and order data set, analyze the difference between the automotive modification parts and the existing products, judge the complexity level of the manufacturer producing the automotive modification parts order, and whether the production process of the existing products needs to be adjusted, then select the production line that meets the demand of the automotive modification parts order from the supplier ranking database GK and the manufacturer ranking database ZK, and generate the corresponding production line data set and custom time Dzsc;

[0013] Step five: Set a fixed range of quantity threshold SLY, and then select the appropriate number of production lines to start the automotive modification parts order combined with the order data set and the production line data set, and monitor the production of each production line during the custom time Dzsc;

[0014] Step six: Set a fixed range of risk fluctuation threshold FXY and quality fluctuation threshold ZLY, and then judge whether the actual production meets the demand of the automotive modification parts order combined with the production of each production line during the custom time Dzsc, and take corresponding optimization measures.

[0015] Preferably, in step one, the expression of the supply data set is {G1 t , G2 t , G3 t ,..., Gy t}, G1 t to Gy t are the management data of the first to the yth supplier, and the management data of the supplier includes part name, part quantity and unqualified quantity, t represents the delivery time of a single supplier, the expression of the manufacturing data set is {Z1 s , Z2 s , Z3 s ,..., Zh s}, Z1 s to Zh sThe management data of the first to the hth manufacturers, respectively, includes coaxiality, fatigue life, dimensional tolerance, process step number, and yield rate, s represents the length of the work period of a single manufacturer, and the expression of the order data set is {D1 j , D2 j , D3 j ,..., Dn j}, D1 j to Dn j are the demand data of the first to the nth automobile modification part orders, respectively, and the management data of the automobile modification part orders includes modification type, size area, material hardness, color value, and customization quantity, and j represents the delivery length of a single modification part order.

[0016] Preferably, in step two, the risk value Fxz is calculated according to the following procedure:

[0017] S11, extract the management data of the e th supplier in the supply data set, and mark the part name provided by the e th supplier as LM e , mark the part quantity provided by the e th supplier as LS e , mark the unqualified quantity of the e th supplier as LC e , and mark the delivery length of the e th supplier as e t ;

[0018] S12, when a batch of parts LM e is delivered, the delivery rate JV e of the e th supplier is calculated, and the calculation formula is as follows:

[0019]

[0020] S13, set a fixed-length monitoring period Q, and statistically analyze the change trend of the delivery rate JV e of the e th supplier during the monitoring period Q, and generate the corresponding supply coefficient Gyxs e , and the calculation formula is as follows:

[0021]

[0022] In the formula, ZJ represents the total delivery times of the e th supplier during the monitoring period Q, represents the delivery rate of the e th supplier when the u th batch of parts is delivered during the monitoring period Q, represents the average delivery rate of the e th supplier during the monitoring period Q, and the variance value obtained according to the variance formula is the supply coefficient Gyxs e of the e th supplier.

[0023] S14, calculate the risk value Fxz of the e-th supplier according to S11-S13 e The calculation formula is as follows:

[0024]

[0025] In the formula, α1 represents the evaluation weight of the supply coefficient, represents the ratio of the unqualified quantity to the part quantity, i.e. the unqualified rate of the e-th supplier, α2 represents the evaluation weight of the unqualified rate, and α1+α2=1, represents the risk value Fxz of the e-th supplier obtained by comprehensively considering the supply coefficient and the unqualified rate according to the weights α1 and α2. The risk value Fxz will be recalculated when the single monitoring period Q ends.

[0026] Preferably, in step two, the supplier ranking database GK establishes corresponding partitions according to different part names, and stores the management data of the suppliers from low to high in each partition according to the risk value Fxz.

[0027] Preferably, in step three, the quality coefficient Zlxs is calculated according to the following process:

[0028] S21, extract the management data of the f-th manufacturer in the manufacturing data set, and mark the concentricity of the f-th manufacturer during production as TZ f mark the fatigue life of the f-th manufacturer during production as PS f mark the dimensional tolerance of the f-th manufacturer during production as CU f mark the process step number of the f-th manufacturer during production as YB f mark the yield rate of the f-th manufacturer during production as YL f ;

[0029] S22, calculate the assembly accuracy Zujd of the f-th manufacturer every time a batch of products is produced f The calculation formula is as follows:

[0030] Zujd f = ω1×TZ f + ω2×PS f - ω3×CU f ;

[0031] In the formula, ω1 represents the evaluation weight of the concentricity, ω2 represents the evaluation weight of the fatigue life, and ω3 represents the evaluation weight of the dimensional tolerance, ω1+ω2+ω3=1, ω1×TZ f + ω2×PS f - ω3×CU fZujd represents the assembly accuracy of the fth manufacturer, which is obtained by integrating the coaxiality, fatigue life and dimensional tolerance according to the weights of ω1, ω2 and ω3 f ;

[0032] S23, integrating S21-S22, calculate the quality coefficient Zlxs of the fth manufacturer f , the calculation formula is as follows:

[0033]

[0034] In the formula, represents the evaluation weight for assembly accuracy, represents the ratio of yield rate to process steps, represents the evaluation weight for the ratio of yield rate to process steps, represents the quality coefficient Zlxs of the fth manufacturer obtained by integrating the assembly accuracy, yield rate and process steps according to the weights of and f The update frequency of the assembly accuracy Zujd and the quality coefficient Zlxs of the single manufacturer is consistent.

[0035] Preferably, in step three, the manufacturer ranking database ZK stores the management data of the manufacturers from high to low according to the quality coefficient Zlxs.

[0036] Preferably, in step four, the customization time Dzsc calculation process is as follows:

[0037] S31, extract the demand data of the gth automobile modification part order in the order data set, and mark the modification type of the gth automobile modification part order as GL g Mark the size area of the gth automobile modification part order as DC g Mark the material hardness of the gth automobile modification part order as GY g Mark the color value of the gth automobile modification part order as SD g Mark the customization quantity of the gth automobile modification part order as DS g Mark the delivery time of the gth automobile modification part order as g j ;

[0038] S32, calculate the difference data group CY of the gth automobile modification part order g , the calculation formula is as follows:

[0039]

[0040] In the formula, BCH represents the standard size area when the manufacturer produces existing products, DC​g -BCH represents the difference between the size area of the gth order of automobile modification parts and the standard size area, BYD represents the standard material hardness when the manufacturer produces existing products, GY g -BYD represents the difference between the material hardness of the gth order of automobile modification parts and the standard material hardness, BSD represents the standard colorimetric value when the manufacturer produces existing products, SD g -BSD represents the difference between the colorimetric value of the gth order of automobile modification parts and the standard colorimetric value;

[0041] S33, according to the difference data set CY g of the gth order of automobile modification parts, determine the complexity level of the manufacturer producing the gth order of automobile modification parts, and the first complexity level is higher than the second complexity level;

[0042] In the first case, any value in the difference data set CY g is not equal to 0, indicating that the manufacturer needs to adjust the production process of existing products when producing the gth order of automobile modification parts, generating a first complexity level, and according to the modification type GL g of the gth order of automobile modification parts, setting the delay parameter μ and generating the corresponding production line data set, wherein the delay parameter of performance modification is 5 days, the delay parameter of appearance modification is 4 days, the delay parameter of interior modification is 3 days, the delay parameter of functional modification is 4 days, and the delay parameter of electronic modification is 3 days;

[0043] In the second case, all values in the difference data set CY g are 0, indicating that the manufacturer does not need to adjust the production process of the manufacturer when producing the gth order of modification parts, generating a second complexity level, and directly generating the corresponding production line data set;

[0044] S34, according to the management data of the gth order of automobile modification parts, set the constraint conditions, the constraint conditions include risk value Fxz, quality coefficient Zlxs, delivery time length, construction period time length and logistics distance, in the constraint conditions, the priority of risk value Fxz and quality coefficient Zlxs is the highest, the priority of delivery time length and construction period time length is higher than that of logistics distance, then from the supplier ranking database GK and the manufacturer ranking database ZK, filter out the production line that meets the demand of the gth order of automobile modification parts, generate the corresponding production line data set, the production line data set includes multiple production lines, and each production line includes multiple suppliers and a single manufacturer;

[0045] S35, in the first case, when generating the first complexity level, comprehensively S33-S34, calculate the customization time length Dzsc g required by the manufacturer to produce the gth order of automobile modification parts, and the calculation formula is as follows:

[0046] Cxa = {Gi t ,...,Zk s}, i∈y, k∈h;

[0047]

[0048] In the second case, when generating the first complexity level, the customized duration Dzsc required for the manufacturer to produce the gth order of the automobile modification part is calculated by integrating S33-S34. g The calculation formula is as follows:

[0049]

[0050] In the formula, Cx a represents the ath production line in the production line data set, including the management data of multiple suppliers and a single manufacturer, b represents the total number of production lines in the production line data set, θ represents a conversion coefficient, and θxCx a represents the conversion of the delivery duration t of the supplier and the construction duration s of the manufacturer in the ath production line into the corresponding customized duration.

[0051] Preferably, in the fifth step, if the delivery duration g j of the gth order of the automobile modification part is less than the customized duration Dzsc g , and DS g is lower than or included in the quantity threshold SLY, only one production line in the production line data set is selected to start the automobile modification part order, if the delivery duration g j of the gth order of the automobile modification part is greater than or equal to the customized duration Dzsc g , and DS g exceeds the quantity threshold SLY, two production lines in the production line data set are selected to start the automobile modification part order simultaneously, if the delivery duration g j of the gth order of the automobile modification part is greater than or equal to the customized duration Dzsc g , and DS g exceeds twice the quantity threshold SLY, three production lines in the production line data set are selected to start the automobile modification part order simultaneously.

[0052] Preferably, in the sixth step, if the variation of the single supplier risk value Fxz during the customized duration Dzsc exceeds the risk fluctuation threshold FXY or the variation of the single manufacturer quality coefficient Zlxs exceeds the quality fluctuation threshold ZLY, it indicates that the actual production situation does not meet the requirements of the automobile modification part order, and new suppliers or manufacturers need to be selected from the supplier ranking database GK and the manufacturer ranking database ZK to replace the original ones.

[0053] The supply chain optimization decision system based on artificial intelligence comprises a customized service module, a data analysis module, an intelligent evaluation module and an optimization decision module;

[0054] The customized service module connects a supply chain management system through a network, obtains management data of all suppliers and manufacturers and demand data of all automobile modification part orders in real time, and classifies and forms a supply data set, a manufacturing data set and an order data set;

[0055] The data analysis module generates corresponding supply coefficients Gyxs and risk values Fxz according to the supply data set, and forms a supplier ranking database GK, and analyzes assembly precision Zujd and quality coefficients Zlxs of each manufacturer according to the manufacturing data set, and forms a manufacturer ranking database ZK;

[0056] The intelligent evaluation module judges the complexity level of the manufacturers in producing the automobile modification part orders according to the supply data set, the manufacturing data set and the order data set, selects line data groups meeting the demand of the automobile modification part orders from the supplier ranking database GK and the manufacturer ranking database ZK, and calculates a customized duration Dzsc;

[0057] The optimization decision module is provided with a fixed range of quantity threshold SLY, a risk fluctuation threshold FXY and a quality fluctuation threshold ZLY, selects a proper number of production lines to start the automobile modification part orders, monitors the production of each production line during the customized duration Dzsc, judges whether the actual production meets the demand of the automobile modification part orders, and carries out corresponding optimization measures.

[0058] Compared with the prior art, the method and system for supply chain optimization decision based on artificial intelligence have the following beneficial effects:

[0059] 1. The application realizes real-time acquisition of management data of all suppliers and manufacturers and demand data of all automobile modification order through the customized service module, and classifies and forms supply data set, manufacturing data set and order data set, and the data analysis module analyzes the Gyxs and risk value Fxz of each supplier and the assembly accuracy Zujd and quality coefficient Zlxs of each manufacturer according to the supply data set and the manufacturing data set, and forms the supplier ranking database GK and the manufacturer ranking database ZK, so that the update frequency of the database is consistent with the information change frequency of each node in the supply chain, the intelligent evaluation module analyzes the difference between the automobile modification and the existing product, judges the complexity level of the manufacturer producing the automobile modification order and whether the production process of the existing product needs to be adjusted, and then selects the production line meeting the demand of the automobile modification order from the supplier ranking database GK and the manufacturer ranking database ZK, generates the corresponding production line data group and the customization time Dzsc, and when the existing production line cannot meet the complex demand of the automobile modification, the most suitable production line is matched from the supplier ranking database GK and the manufacturer ranking database ZK, so that the artificial judgment and the delay of the order are reduced, the response speed and the efficiency of the whole production process are improved, and the intelligent configuration production line has high flexibility.

[0060] 2. The application selects a proper number of production lines to start the automobile modification order through the optimization decision module, monitors the production of each production line during the customization time Dzsc, judges whether the actual production meets the demand of the automobile modification order, and takes corresponding optimization measures, if the change amount of the risk value Fxz of a single supplier or the change amount of the quality coefficient Zlxs of a single manufacturer exceeds the risk fluctuation threshold FXY or the quality fluctuation threshold ZLY during the customization time Dzsc, the actual production does not meet the demand of the automobile modification order, which may cause the quality of the modification part not meeting the order demand or the delivery time being delayed, so that new suppliers or manufacturers are selected from the supplier ranking database GK and the manufacturer ranking database ZK to replace the old ones, and the product quality is more stable after dynamic adjustment. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 The method steps of the application are shown in the figure;

[0062] Figure 2 The system flowchart of the application is shown in the figure. DETAILED DESCRIPTION

[0063] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0064] Embodiment 1

[0065] Please refer to Figure 1 The present application provides the method for optimizing the supply chain decision based on artificial intelligence, which comprises the following steps:

[0066] Step one: connect the supply chain management system through the network, acquire the management data of all suppliers and manufacturers and the demand data of all automobile modification order in real time, and classify and form the supply data set, the manufacturing data set and the order data set;

[0067] Step two: according to the supply data set, analyze the delivery rate JV, the change trend of the delivery rate JV and the delivery risk of each supplier, generate the corresponding supply coefficient Gyxs and risk value Fxz, and form the supplier ranking database GK;

[0068] Step three: according to the manufacturing data set, analyze the assembly accuracy Zujd and the quality coefficient Zlxs of each manufacturer, and form the manufacturer ranking database ZK;

[0069] Step four: according to the supply data set, the manufacturing data set and the order data set, analyze the difference between the automobile modification and the existing product, judge the complexity level of the manufacturer producing the automobile modification order and whether the production process of the existing product needs to be adjusted, then select the production line meeting the demand of the automobile modification order from the supplier ranking database GK and the manufacturer ranking database ZK, and generate the corresponding production line data set and custom time Dzsc;

[0070] Step five: set a fixed range of quantity threshold SLY, then combine the order data set and the production line data set, select a proper number of production lines to start the automobile modification order, and monitor the production of each production line during the custom time Dzsc;

[0071] Step six: set a fixed range of risk fluctuation threshold FXY and quality fluctuation threshold ZLY, then combine the production of each production line during the custom time Dzsc, judge whether the actual production meets the demand of the automobile modification order, and take corresponding optimization measures.

[0072] In this embodiment, SQL query technology is used to retrieve management data from all suppliers and manufacturers, as well as demand data from all automotive modification parts orders, in real time from the supply chain management system. This data is then categorized into supply datasets, manufacturing datasets, and order datasets. This helps artificial intelligence retrieve the required content more quickly and accurately when searching information from multiple sources. Based on the supply dataset, the delivery rate JV, its changing trend, and delivery risk for each supplier are analyzed to generate corresponding supply coefficients Gyxs and risk values ​​Fxz, forming a supplier ranking database GK. Based on the manufacturing dataset, the assembly accuracy Zujd and quality coefficient Zlxs for each manufacturer are analyzed, forming a manufacturer ranking database ZK. Two databases with ranking mechanisms are established based on the supply chain management system. When information and demand at each node in the supply chain change rapidly, the update frequency of the supplier ranking database GK and the manufacturer ranking database ZK can also be synchronized with these node changes. This process analyzes the differences between automotive modification parts and existing products to determine the complexity level of automotive modification part orders and whether adjustments to existing product manufacturing processes are necessary. Then, it filters the supplier ranking database (GK) and manufacturer ranking database (ZK) to select production lines that meet the requirements of automotive modification part orders, generating corresponding production line data groups and customization durations (Dzsc). When existing production lines cannot meet the complex requirements of automotive modification parts, the most suitable production line is automatically matched from the supplier ranking database (GK) and manufacturer ranking database (ZK), reducing manual judgment and order delays, and improving the responsiveness and efficiency of the entire production process. By selecting an appropriate number of production lines to initiate automotive modification part orders and monitoring the production status of each production line during the customization duration (Dzsc), it determines whether the actual production situation meets the requirements of automotive modification part orders and implements corresponding optimization measures, achieving optimal allocation of supply chain resources and effectively avoiding product quality issues.

[0073] Example 2

[0074] This embodiment is an explanation of Embodiment 1. Specifically, it involves denoising, classifying, and labeling the raw data in the supply chain management system. The denoising process avoids duplicate information in supplier, manufacturer, and auto modification parts orders. The data is then classified and labeled according to the supplier company name. The expression for the supply dataset is {G1}. t G2 t G3 t ... Gy t}, G1 t To Gy t Let Z1 represent the management data for the first to the y-th suppliers. Each supplier's management data includes part name, part quantity, and number of non-conforming parts. Let t represent the delivery time for a single supplier. The management data is categorized and labeled according to the manufacturer's company name. The expression for the manufacturing dataset is {Z1}.s , Z2 s , Z3 s , …, Zh s}, Z1 s to Zh s are the management data of the first to the hth manufacturer respectively, and each manufacturer's management data includes coaxiality, fatigue life, dimensional tolerance, process step number and yield rate, s represents the length of the work period of a single manufacturer, and the demand data is arranged from early to late and sequentially marked according to the ordering time point of the automobile modification order, and the expression of the order data set is {D1 j , D2 j , D3 j , …, Dn j}, D1 j to Dn j are the demand data of the first to the nth automobile modification order respectively, and each automobile modification order management data includes modification type, size area, material hardness, color value and customization quantity, and j represents the delivery length of a single modification order. The supply data set and the manufacturing data set after classification marking represent the supply end in the market supply and demand relationship, and the order data set represents the demand end in the market supply and demand relationship.

[0075] Embodiment 3

[0076] This embodiment is an explanation and description in embodiment 1, specifically, the risk value Fxz calculation process is as follows:

[0077] S11, extract the management data of the e th supplier in the supply data set, and mark the part name provided by the e th supplier as LM e , mark the part quantity provided by the e th supplier as LS e , mark the unqualified quantity of the e th supplier as LC e , mark the delivery length of the e th supplier as e t ;

[0078] S12, when delivering a batch of parts LM e , calculate the delivery rate JV e of the e th supplier, and the calculation formula is as follows:

[0079]

[0080] S13, set a fixed length monitoring period Q, and statistically analyze the change trend of the delivery rate JV e of the e th supplier during the monitoring period Q, and generate the corresponding supply coefficient Gyxs e , and the calculation formula is as follows:

[0081]

[0082] In the formula, ZJ represents the total number of deliveries by the e-th supplier during the monitoring period Q. This represents the delivery rate of the e-th supplier when delivering the u-th batch of parts during the monitoring period Q. This represents the average delivery rate of the e-th supplier during the monitoring period Q. This means that, according to the variance formula, the variance value obtained is the supply coefficient Gyxs of the e-th supplier. e ;

[0083] S14. Combining S11-S13, calculate the risk value Fxz of the e-th supplier. e The calculation formula is as follows:

[0084]

[0085] In the formula, α1 represents the evaluation weight for the supply coefficient. The ratio of the number of defective parts to the total number of parts is the defect rate of the e-th supplier. α2 represents the evaluation weight for the defect rate, and α1 + α2 = 1. This means that the risk value Fxz of the e-th supplier is obtained by combining the supply coefficient and the non-conformance rate according to the weights of α1 and α2. The risk value Fxz will be recalculated at the end of a single monitoring period Q.

[0086] Specific numerical examples:

[0087] The supplier of wheel hub bearings was selected as the experimental subject. The calculation process for the supplier's risk value Fxz is as follows:

[0088] S11. The supplier provides 100 parts, has 3 defective parts, and has a delivery time of 5 days.

[0089] S12. For each batch of wheel hub bearings delivered, the supplier's delivery rate JV e The calculation formula is as follows:

[0090]

[0091] S13. Set a fixed monitoring period Q of 20 days. During the monitoring period Q, the supplier's delivery rate JV e The supplier's supply coefficient Gyxs is 20 units / day, 18 units / day, 16 units / day, 22 units / day, and 19 units / day respectively. e The calculation formula is as follows:

[0092]

[0093] In the formula, ZJ represents the total delivery times of the supplier during the monitoring period Q, which is 5 times, represents the delivery rate of the supplier when delivering the u-th batch of parts during the monitoring period Q, which is 20 pieces / day, represents the average delivery rate of the supplier during the monitoring period Q, which is 19 pieces / day, and the variance value 4 obtained according to the variance formula is the supply coefficient Gyxs of the supplier e ;

[0094] S14, comprehensively calculate the risk value Fxz of the supplier according to S11-S13 e , and the calculation formula is as follows:

[0095]

[0096] In the formula, 0.6 represents the evaluation weight for the supply coefficient, represents the ratio of the unqualified quantity to the part quantity, that is, the unqualified rate of the supplier, 0.4 represents the evaluation weight for the unqualified rate, and 0.6+0.4=1, represents that the supply coefficient and the unqualified rate are comprehensively calculated according to the weights of a1 and a2, and the risk value Fxz of the supplier is 2.412, and the risk value Fxz will be recalculated when each monitoring period Q ends;

[0097] In this embodiment, each supplier has a corresponding risk value Fxz, and the higher the risk value Fxz, the worse the stability of the supplier in normal delivery of parts. The supplier ranking database GK establishes corresponding partitions according to different part names, for example, hub bearing partition, hub cover partition, hub bolt partition, etc. In each partition, the ranking mechanism stores the management data of the supplier from low to high according to the risk value Fxz.

[0098] Embodiment 4

[0099] This embodiment is an explanation and description in embodiment 3. Specifically, the calculation process of the quality coefficient Zlxs is as follows:

[0100] S21, extract the management data of the f-th manufacturer in the manufacturing data set, and mark the coaxiality when the f-th manufacturer produces as TZ f , mark the fatigue life when the f-th manufacturer produces as PS f , mark the size tolerance when the f-th manufacturer produces as CU f , mark the process step number when the f-th manufacturer produces as YB f , mark the yield when the f-th manufacturer produces as YL f ;

[0101] S22, calculate the assembly accuracy Zujd of the fth manufacturer every time a batch of products is produced f , whose calculation formula is as follows:

[0102] Zujd f = ω1 × TZ f + ω2 × PS f - ω3 × CU f ;

[0103] In the formula, ω1 represents the evaluation weight for coaxiality, ω2 represents the evaluation weight for fatigue life, and ω3 represents the evaluation weight for dimensional tolerance, ω1 + ω2 + ω3 = 1, ω1 × TZ f + ω2 × PS f - ω3 × CU f represents the assembly accuracy Zujd of the fth manufacturer obtained by integrating coaxiality, fatigue life, and dimensional tolerance according to the weights of ω1, ω2, and ω3 f ;

[0104] S23, calculate the quality coefficient Zlxs of the fth manufacturer by integrating S21-S22 f , whose calculation formula is as follows:

[0105]

[0106] In the formula, represents the evaluation weight for assembly accuracy, represents the ratio of yield rate to process steps, represents the evaluation weight for the ratio of yield rate to process steps, represents the quality coefficient Zlxs of the fth manufacturer obtained by integrating assembly accuracy, yield rate, and process steps according to the weights of and ; f The update frequency of the assembly accuracy Zujd and the quality coefficient Zlxs of a single manufacturer is consistent;

[0107] Specific numerical examples:

[0108] Select the manufacturer of assembling hubs as the experimental object, and the calculation process of the quality coefficient Zlxs of the manufacturer is as follows:

[0109] S21, the coaxiality of the manufacturer assembling hubs is 0.05 mm, the fatigue life of the manufacturer assembling hubs is 50,000 kilometers, the dimensional tolerance of the manufacturer assembling hubs is 0.02 mm, the process steps of the manufacturer assembling hubs are 5, and the yield rate of the manufacturer assembling hubs is 95%;

[0110] S22, calculate the assembly precision Zujd of the manufacturer for each batch of products produced f , and the calculation formula is as follows:

[0111] Zujd f = 0.3 x 0.05 + 0.4 x 5 - 0.3 x 0.02 = 1.999;

[0112] In the formula, 0.3 represents the evaluation weight for coaxiality, 0.4 represents the evaluation weight for fatigue life, and 0.3 represents the evaluation weight for dimensional tolerance, 0.3 + 0.4 + 0.3 = 1, ω1 x TZ f + ω2 x PS f - ω3 x CU f represents the assembly precision Zujd of the manufacturer according to the weights of ω1, ω2 and ω3, which integrates coaxiality, fatigue life and dimensional tolerance f , and the value is 1.999;

[0113] S23, calculate the quality coefficient Zlxs of the manufacturer by integrating S21-S22 f , and the calculation formula is as follows:

[0114]

[0115] In the formula, 0.6 represents the evaluation weight for assembly precision, represents the ratio of yield rate to process steps, 0.4 represents the evaluation weight for the ratio of yield rate to process steps, 0.6 + 0.4 = 1, represents the quality coefficient Zlxs of the manufacturer according to the weights of and , which integrates assembly precision, yield rate and process steps f , and the value is 1.2754. The update frequency of the assembly precision Zujd and the quality coefficient Zlxs of the single manufacturer is consistent;

[0116] In this embodiment, each manufacturer has a corresponding quality coefficient Zlxs, and the higher the quality coefficient Zlxs, the better the quality of the products produced by the manufacturer. The manufacturer ranking database ZK stores the management data of the manufacturers from high to low according to the quality coefficient Zlxs;

[0117] Embodiment 5

[0118] This embodiment is an explanation and description in embodiment 4, and specifically, the customized time length Dzsc calculation process is as follows:

[0119] S31, extract the demand data of the gth automobile modification order in the order data set, and mark the modification type of the gth automobile modification order as GL gThe size and area of ​​the g-th automotive modification parts order will be marked as DC. g The material hardness of the gth automotive modification parts order will be marked as GY. g Mark the chromaticity value of the g-th automotive modification parts order as SD. g The customization quantity of the g-th car modification parts order is marked as DS. g The delivery time of the gth automotive modification parts order is marked as g. j ;

[0120] S32. Calculate the difference data group CY for the g-th automotive modification parts order. g The calculation formula is as follows:

[0121]

[0122] In the formula, BCH represents the standard size area when the manufacturer produces existing products, and DC... g -BCH represents the difference between the area of ​​the g-th automotive modification part order and the standard area; BYD represents the standard material hardness used by the manufacturer when producing existing products; GY g -BYD represents the difference between the material hardness of the g-th automotive modification part order and the standard material hardness; BSD represents the standard colorimetric value used by the manufacturer when producing existing products; SD... g -BSD represents the difference between the chromaticity value of the g-th automotive modification parts order and the standard chromaticity value;

[0123] S33, Based on the difference data group CY of the g-th automotive modification parts order g Determine the complexity level when the manufacturer produces the g-th automotive modification parts order, with the first complexity level being higher than the second complexity level;

[0124] In the first case, the differential data group CY g If any value in the table is not equal to 0, it means that when the manufacturer produces the g-th automotive modification part order, it needs to adjust the production process of the existing products to generate the first complexity level, based on the modification type GL of the g-th automotive modification part order. g While setting the delay parameter ω, the corresponding production line data group is generated. The delay parameter for performance modifications is 5 days, the delay parameter for appearance modifications is 4 days, the delay parameter for interior modifications is 3 days, the delay parameter for functional modifications is 4 days, and the delay parameter for electronic modifications is 3 days.

[0125] In the second case, the differential data group CY g If all values ​​are 0, it means that when the manufacturer produces the g-th modification part order, there is no need to adjust the manufacturer's production process to generate the second level of complexity; the corresponding production line data group is generated directly.

[0126] S34, according to the management data of the gth automobile modification order, set the constraint condition, the constraint condition includes the risk value Fxz, the quality coefficient Zlxs, the delivery time length, the construction period time length and the logistics distance, in the constraint condition, the priority of the risk value Fxz and the quality coefficient Zlxs is highest, the priority of the delivery time length and the construction period time length is higher than the logistics distance, and then the production line meeting the demand of the gth automobile modification order is screened out from the supplier ranking database GK and the manufacturer ranking database ZK, and the corresponding production line data group is generated, the production line data group includes a plurality of production lines, and each production line includes a plurality of suppliers and a single manufacturer;

[0127] S35, in the first case, when the first complexity level is generated, the customized time length Dzsc required for the manufacturer to produce the gth automobile modification order is calculated by comprehensively considering S33-S34 g , and the calculation formula is as follows:

[0128] Cx a = {Gi t ,...,Zk s}, i ∈ y, k ∈ h;

[0129]

[0130] In the second case, when the first complexity level is generated, the customized time length Dzsc required for the manufacturer to produce the gth automobile modification order is calculated by comprehensively considering S33-S34 g , and the calculation formula is as follows:

[0131]

[0132] In the formula, Cx a represents the management data of the a th production line in the production line data group, including a plurality of suppliers and a single manufacturer, b represents the total number of production lines in the production line data group, θ represents the conversion coefficient, and θ × Cx a represents the conversion of the delivery time length t of the supplier and the construction period time length s of the manufacturer in the a th production line into the corresponding customized time length.

[0133] Specific numerical examples:

[0134] The hub modification is selected as the experimental object, the hub modification belongs to the performance modification, the calculation process of the production line data group and the customized time length Dzsc of the hub modification order is as follows:

[0135] S31, the size area, material hardness, color value and customization quantity of each hub modification in the automobile modification order are obtained, the delivery time length of the automobile modification order is unified as 5 days;

[0136] S32, calculate the difference data group CY of the automobile modification orderg , and the calculation formula is as follows:

[0137]

[0138] In the formula, BCH represents the standard size area of the existing product produced by the manufacturer, that is, 254 in 2 , DC g BCH represents the difference between the size area of the automobile modification order and the standard size area, BYD represents the standard material hardness of the existing product produced by the manufacturer, that is, 90 HB, GY g BYD represents the difference between the material hardness of the automobile modification order and the standard material hardness, and BSD represents the standard chroma value of the existing product produced by the manufacturer, that is, RGB (169, 169, 169), SD g BSD represents the difference between the chroma value of the gth automobile modification order and the standard chroma value.

[0139] S33, referring to Table 1, according to the difference data set CY g of the automobile modification order, the complexity level of the manufacturer producing the gth automobile modification order is determined.

[0140] In the first case, any value in the difference data set CY g is not equal to 0, indicating that the production process of the existing product needs to be adjusted when the manufacturer produces the gth automobile modification order, a first complexity level is generated, the modification type GL g of the automobile modification order corresponds to a delay parameter h of 5 days, and the corresponding production line data set is generated.

[0141] In the second case, all values in the difference data set CY g are 0, indicating that the production process of the manufacturer does not need to be adjusted when the manufacturer produces the gth modification order, a second complexity level is generated, and the corresponding production line data set is directly generated.

[0142]

[0143] Table 1

[0144] S34, please refer to Table 2, according to the management data of the automobile modification order, set constraint conditions, constraint conditions include risk value Fxz, quality coefficient Zlxs, delivery time, construction period and logistics distance, in the constraint condition, the priority of risk value Fxz and quality coefficient Zlxs is the highest, the priority of delivery time and construction period is higher than logistics distance, and then from the supplier ranking database GK and the manufacturer ranking database ZK, the production line meeting the requirements of the gth automobile modification order is screened out, and the corresponding production line data group is generated, the production line data group includes multiple production lines, and each production line includes multiple suppliers and a single manufacturer;

[0145]

[0146] Table 2

[0147] Specifically, when the risk value Fxz and the quality coefficient Zlxs are given priority, the quality coefficient Zlxs of line 4 is the highest, the risk value is lower, and the delivery time and construction period are moderate, line 4 is the best configured production line, when the delivery time and construction period are considered secondly, the quality coefficient of line 1 is higher, the risk value is lower, and the delivery time and construction period are moderate, the defect is that the logistics distance is longer, which is suitable for processing non-urgent orders, when the logistics distance is considered secondly, although the risk value Fxz and the quality coefficient Zlxs of line 2 are relatively low, the delivery time and construction period are short, and the logistics distance is relatively short, which is suitable for processing complex and low-grade urgent orders;

[0148] S35, please refer to Table 3, S33-S34, calculate the customization time Dzsc required for the manufacturer to produce the automobile modification order g , the calculation formula is as follows:

[0149] If the manufacturer produces the automobile modification order, the production process of the existing product needs to be adjusted, the processing complexity is high,

[0150] Cx a ={Gi t ,...,Zk s},i∈y,k∈h;

[0151]

[0152] If the manufacturer produces the modification order, the production process of the manufacturer does not need to be adjusted, the processing complexity is low,

[0153]

[0154] In the formula, Cx aLet represent the 'a'-th production line in the production line data group, including management data from multiple suppliers and a single manufacturer; let 'b' represent the total number of production lines in the production line data group; and let 'θ' represent the conversion coefficient, which is derived from the current logistics delivery speed. The faster the logistics delivery speed, the smaller the conversion coefficient. θ×Cx a This indicates that the delivery time t of the supplier and the lead time s of the manufacturer in the a-th production line are converted into the corresponding customized time.

[0155] When the conversion factor θ at the current time point is 0.8, calculate the customization time for each wheel modification part in this car modification parts order. The customization time for wheel modification parts 2 and 3 requires an additional delay parameter μ, which is 5 days.

[0156]

[0157] Table 3

[0158] Example 6

[0159] This embodiment is an explanation based on Embodiment 5. Specifically, the quantity threshold SLY comes from the assembly limit of a single batch on the production line, the risk fluctuation threshold FXY comes from the company's evaluation criteria for suppliers, and the quality fluctuation threshold ZLY comes from the company's evaluation criteria for manufacturers. If the delivery time of this automotive modification parts order is g... j Less than the customized duration Dzsc g , and DS g When the quantity is below or within the quantity threshold SLY, only one production line in the production line data group is selected to initiate an automotive modification parts order. If the delivery time of the g-th automotive modification parts order is g... j Greater than or equal to the customized duration Dzsc g , and DS g When the quantity exceeds the threshold SLY, select two production lines in the production line data group to simultaneously start the automotive modification parts order. If the delivery time of the g-th automotive modification parts order is g... j Greater than or equal to the customized duration Dzsc g , and DS g When the quantity exceeds twice the threshold SLY, select three production lines in the production line data group to simultaneously start the automotive modification parts order;

[0160] The production line synchronously starts the automobile modification part order, and monitors the production of each production line during the customization time Dzsc through the supply chain management system, judges whether the actual production condition meets the automobile modification part order demand in real time, and carries out corresponding optimization measures. If the change amount of the single supplier risk value Fxz exceeds the risk fluctuation threshold FXY or the change amount of the single manufacturer quality coefficient Zlxs exceeds the quality fluctuation threshold ZLY during the customization time Dzsc, it indicates that the actual production condition does not meet the automobile modification part order demand, and new suppliers or manufacturers need to be selected from the supplier ranking database GK and the manufacturer ranking database ZK to replace them.

[0161] Embodiment 7

[0162] Please refer to Figure 2 The artificial intelligence-based supply chain optimization decision system includes a customization service module, a data analysis module, an intelligent evaluation module, and an optimization decision module.

[0163] The customization service module connects the supply chain management system through the network, obtains the management data of all suppliers and manufacturers and the demand data of all automobile modification part orders in real time, and classifies and forms the supply data set, the manufacturing data set, and the order data set;

[0164] The data analysis module generates the corresponding supply coefficient Gyxs and risk value Fxz according to the supply data set, and forms the supplier ranking database GK. The data analysis module analyzes the assembly accuracy Zujd and the quality coefficient Zlxs of each manufacturer according to the manufacturing data set, and forms the manufacturer ranking database ZK.

[0165] The intelligent evaluation module judges the complexity level of the manufacturer producing the automobile modification part order according to the supply data set, the manufacturing data set, and the order data set, selects the production line data group that meets the automobile modification part order demand from the supplier ranking database GK and the manufacturer ranking database ZK, and calculates the customization time Dzsc.

[0166] The optimization decision module is provided with a fixed range of quantity threshold SLY, risk fluctuation threshold FXY, and quality fluctuation threshold ZLY, selects a proper number of production lines to start the automobile modification part order, monitors the production of each production line during the customization time Dzsc, judges whether the actual production condition meets the automobile modification part order demand, and carries out corresponding optimization measures.

[0167] In this embodiment, the manufacturer ranking database ZK and the manufacturer ranking database ZK are established, which provides more accurate data support for the production line configuration, eliminates the information barrier between the nodes of the supply management system, and has high flexibility of the intelligent configuration production line. According to the supply data set, the manufacturing data set and the order data set, the complexity level of the manufacturer producing the automobile modification part order is judged, and then the production line data set meeting the demand of the automobile modification part order is screened out from the supplier ranking database GK and the manufacturer ranking database ZK, and the customization time Dzsc is calculated. Different types of modification parts correspond to different complexity judgment conditions, for example, the judgment conditions of appearance modification parts are film thickness, surface roughness, etc. The intelligent configuration production line has high flexibility, a proper number of production lines are selected to start the automobile modification part order, the production situation of each production line during the customization time Dzsc is monitored, whether the actual production situation meets the demand of the automobile modification part order is judged, and corresponding optimization measures are taken. When the quality of the modification part does not meet the demand of the order or the delivery time is delayed, the supplier and the manufacturer can be replaced in time, and the product quality is dynamically adjusted to be more stable.

[0168] Although the embodiments of the present application have been shown and described, it is to be understood that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for artificial intelligence based supply chain optimization decisions, characterized in that, The method comprises the following steps: Step one: real-time acquisition of all supplier and manufacturer management data and all automobile modification part order demand data through network connection of the supply chain management system, and classification into a supply data set, a manufacturing data set and an order data set; Step two: analysis of the delivery rate JV, the change trend of the delivery rate JV and the delivery risk of each supplier according to the supply data set, generation of the corresponding supply coefficient Gyxs and risk value Fxz, and formation of a supplier ranking database GK; Step three: analysis of the assembly accuracy Zujd and the quality coefficient Zlxs of each manufacturer according to the manufacturing data set, and formation of a manufacturer ranking database ZK; Step four: According to the supply dataset, manufacturing dataset and order dataset, analyze the differences between the size area, material hardness and color value of the gth automobile modification order and the standard size area, standard material hardness and standard color value when the manufacturer produces the existing product, and generate a difference dataset CY g And based on the difference dataset CY g Judge the complexity level when the manufacturer produces the automobile modification order and whether the production process of the existing product needs to be adjusted, and then filter the production line that meets the demand of the automobile modification order from the supplier ranking database GK and the manufacturer ranking database ZK, and generate the corresponding production line dataset and customization time Dzsc; Step five: setting of an assembly upper limit value SLY of a single batch of production lines, combination of the order data set and the production line data set, selection of a proper number of production lines to start the automobile modification part order, and monitoring of the production of each production line during the customization time length Dzsc; Step six: setting of a risk fluctuation threshold FXY and a quality fluctuation threshold ZLY, combination of the production of each production line during the customization time length Dzsc, judgment of whether the actual production meets the automobile modification part order demand, and corresponding optimization measures.

2. The method of artificial intelligence based supply chain optimization decision making as claimed in claim 1 wherein: The expression of the supply data set in step one is {G1 t , G2 t , G3 t ,..., Gy t}, G1 t , G2 t , G3 t ,..., Gy t are the management data of the first supplier to the yth supplier respectively, the management data of the supplier includes the part name, the part quantity and the unqualified quantity, t represents the delivery time length of a single supplier, the expression of the manufacturing data set is {Z1 s , Z2 s , Z3 s ,..., Zh s}, Z1 s , Z2 s , Z3 s ,..., Zh s are the management data of the first manufacturer to the hth manufacturer respectively, the management data of the manufacturer includes the coaxiality, the fatigue life, the size tolerance, the process step number and the yield rate, s represents the construction period length of a single manufacturer, the expression of the order data set is {D1 j , D2 j , D3 j ,..., Dn j}, D1 j , D2 j , D3 j ,..., Dn j are the demand data of the first automobile modification part order to the nth automobile modification part order respectively, the demand data of the automobile modification part order includes the modification type, the size area, the material hardness, the color value and the customized quantity, j represents the delivery time length of a single modification part order.

3. The method of AI-based supply chain optimization decision making as claimed in claim 2, wherein: In the step two, the risk value Fxz calculation process is as follows: S11, extract management data of the e-th supplier in the supply data set, and mark the part name provided by the e-th supplier as LM e mark the part quantity provided by the e-th supplier as LS e mark the unqualified quantity of the e-th supplier as LC e mark the delivery time of the e-th supplier as e t ; S12, for each delivery of a batch of parts LM e the delivery rate JV of the e-th supplier is calculated e with the following formula: S13, set a fixed monitoring period Q, count and analyze the change trend of the delivery rate JV of the e-th supplier during the monitoring period Q, and generate the corresponding supply coefficient Gyxs e e The calculation formula is as follows:​ In the formula, ZJ represents the total delivery times of the e-th supplier during the monitoring period Q, represents the delivery rate of the e-th supplier when delivering the u-th batch of parts during the monitoring period Q, represents the average delivery rate of the e-th supplier during the monitoring period Q, represents the variance value obtained according to the variance formula, which is the supply coefficient Gyxs of the e-th supplier e ; S14, according to S11-S13, the risk value of the e-th supplier Fxz is calculated e The calculation formula is as follows: In the formula, a1 represents an evaluation weight for the supply coefficient, represents the ratio of the number of unqualified products to the number of parts, i.e. the unqualified rate of the e-th supplier, a2 represents an evaluation weight for the unqualified rate, and a1+a2=1, represents the risk value Fxz of the e-th supplier obtained by comprehensively considering the supply coefficient and the unqualified rate according to the weights a1 and a2 e At the end of a single monitoring period Q, the risk value Fxz will be recalculated e .

4. The method of claim 3, wherein: In the step two, the supplier ranking database GK establishes corresponding partitions according to different part names, and stores the management data of the suppliers in each partition from low to high according to the risk value Fxz.

5. The method of claim 4, wherein: In the step three, the quality coefficient Zlxs calculation process is as follows: S21, extract management data of the fth manufacturer in the manufacturing data set, and mark the coaxiality at the time of production of the fth manufacturer as TZ f mark the fatigue life at the time of production of the fth manufacturer as PS f mark the dimensional tolerance at the time of production of the fth manufacturer as CU f mark the process step number at the time of production of the fth manufacturer as YB f mark the yield at the time of production of the fth manufacturer as YL f ; S22, when a batch of products is produced, calculating the assembly precision Zujd of the fth manufacturer f The calculation formula is as follows: Zujd f = ω1 x TZ f + ω2 x PS f - ω3 x CU f ; In the formula, ω1 represents an evaluation weight for coaxiality, ω2 represents an evaluation weight for fatigue life, and ω3 represents an evaluation weight for dimensional tolerance, ω1+ω2+ω3=1, ω1×TZ f +ω2×PS f -ω3×CU f represents an assembly accuracy of the fth manufacturer Zujd obtained by integrating coaxiality, fatigue life, and dimensional tolerance according to the ω1, ω2, and ω3 weights f ; S23、comprehensive S21-S22, the quality coefficient of the fth manufacturer Zlxs is calculated f The calculation formula is as follows: In the formula, represents the evaluation weight for assembly accuracy, represents the ratio of yield rate to process steps, represents the evaluation weight for the ratio of yield rate to process steps, represents the weight according to and the assembly accuracy, the yield rate and the process steps, to obtain the quality coefficient Zlxs of the fth manufacturer f The updating frequency of the assembly accuracy Zujd and the quality coefficient Zlxs of the single manufacturer is consistent.

6. The method of claim 3, wherein: In the step three, the manufacturer ranking database ZK stores the management data of the manufacturers from high to low according to the quality coefficient Zlxs.

7. The method of claim 3, wherein: In the step four, the customization time length Dzsc calculation process is as follows: S31, extract the demand data of the gth automobile modification order in the order data set, and mark the modification type of the gth automobile modification order as GL g mark the size area of the gth automobile modification order as DC g mark the material hardness of the gth automobile modification order as GY g mark the chroma value of the gth automobile modification order as SD g mark the customization quantity of the gth automobile modification order as DS g mark the delivery time length of the gth automobile modification order as g j ; S32, difference data set CY g The calculation formula is as follows: In the formula, BCH represents the standard size area of the existing product produced by the manufacturer, DC g BCH represents the difference between the size area of the gth order of automobile modification parts and the standard size area, BYD represents the standard material hardness of the existing product produced by the manufacturer, GY g BYD represents the difference between the material hardness of the gth order of automobile modification parts and the standard material hardness, BSD represents the standard chroma value of the existing product produced by the manufacturer, SD g BSD represents the difference between the chroma value of the gth order of automobile modification parts and the standard chroma value. S33, if the difference data set CY g Any of the numerical values not equal to 0, indicates that the manufacturer needs to adjust the production process of the existing product when producing the gth order of automobile modification parts, to generate the first complexity level, according to the modification type GL of the gth order of automobile modification parts g , set the delay parameter μ, and at the same time, generate the corresponding production line data set, wherein the delay parameter of the performance class modification is 5 days, the delay parameter of the appearance class modification is 4 days, the delay parameter of the interior class modification is 3 days, the delay parameter of the functional modification is 4 days, and the delay parameter of the electronic class modification is 3 days; If the difference data set CY g All the values in the data set are 0, indicating that when the manufacturer produces the gth order of the modified part, the manufacturing process of the manufacturer does not need to be adjusted, the second complexity level is generated, and the corresponding production line data set is directly generated. S34, setting of a constraint condition according to the demand data of the gth automobile modification part order, the constraint condition including the risk value Fxz, the quality coefficient Zlxs, the delivery time length of the supplier, the construction period time length and the logistics distance, the priority of the risk value Fxz and the quality coefficient Zlxs being the highest, the priority of the delivery time length of the supplier and the construction period time length being higher than that of the logistics distance, screening of a production line meeting the demand of the gth automobile modification part order from the supplier ranking database GK and the manufacturer ranking database ZK, generation of a corresponding production line data set, and the production line data set including multiple production lines and each production line including multiple suppliers and a single manufacturer; S35、in the first g automobile modification parts order is judged to be the first complex level, comprehensive S33-S34, calculate the custom time Dzsc required for the manufacturer to produce the first g automobile modification parts order g The formula is as follows: Cx a = {Gi t ,...,Zk s}, i e {1,2,3,...,y}, k e {1,2,3,...,h}; When the gth order of automobile modification parts is determined as the second complexity level, the customized duration Dzsc required for the manufacturer to produce the gth order of automobile modification parts is calculated according to S33-S34 g The calculation formula is as follows: In the formula, Cx a represents the a-th production line in the production line data set, including management data of multiple suppliers and a single manufacturer, b represents the total number of production lines in the production line data set, θ represents a conversion coefficient, and θ×Cx a represents the conversion of the delivery time t of the supplier and the working period time s of the manufacturer in the a-th production line into the corresponding customization time.

8. The method of claim 7, wherein: In the step six, if the change amount of the risk value Fxz of a single supplier or the change amount of the quality coefficient Zlxs of a single manufacturer exceeds the risk fluctuation threshold FXY or the quality fluctuation threshold ZLY during the customization time length Dzsc, it indicates that the actual production does not meet the automobile modification part order demand, and new suppliers or manufacturers need to be selected from the supplier ranking database GK and the manufacturer ranking database ZK for replacement.

9. The system for artificial intelligence-based supply chain optimization decision-making according to any one of claims 1-8, characterized in that: The system comprises a customization service module, a data analysis module, an intelligent evaluation module and an optimization decision module. The customized service module is connected with a supply chain management system through a network, and acquires management data of all suppliers and manufacturers and demand data of all automobile modification part orders in real time, and classifies and forms a supply data set, a manufacturing data set and an order data set; The data analysis module generates corresponding supply coefficients Gyxs and risk values Fxz according to the supply data set, and forms a supplier ranking database GK, and analyzes assembly precision Zujd and quality coefficients Zlxs of each manufacturer according to the manufacturing data set, and forms a manufacturer ranking database ZK; The intelligent evaluation module judges a complexity level of the manufacturers in producing the automobile modification part orders according to the supply data set, the manufacturing data set and the order data set, selects a production line data group meeting the demand of the automobile modification part orders from the supplier ranking database GK and the manufacturer ranking database ZK, and calculates a customization duration Dzsc; The optimization decision module is provided with a fixed range of a quantity threshold SLY, a risk fluctuation threshold FXY and a quality fluctuation threshold ZLY, selects a proper number of production lines to start the automobile modification part orders, monitors production conditions of each production line during the customization duration Dzsc, judges whether the actual production conditions meet the demand of the automobile modification part orders, and takes corresponding optimization measures.

Images