CN112101769A

CN112101769A - Supply chain risk management system

Info

Publication number: CN112101769A
Application number: CN202010940304.1A
Authority: CN
Inventors: 宋杰; 李银胜
Original assignee: Fudan University
Current assignee: Fudan University
Priority date: 2020-09-09
Filing date: 2020-09-09
Publication date: 2020-12-18

Abstract

The invention provides a supply chain risk management system, which is characterized by comprising the following components: the influence module obtains influence of a single main body and influence among the main bodies by using a knowledge graph, a related graph algorithm and an influence rule; the risk value module analyzes and calculates the data to be analyzed and the influence among the main bodies according to the risk value calculation rule to obtain a main body risk value of each main body after risk association; the safety index module calculates a supply chain safety index corresponding to a plurality of preset safety elements of the supply chain through a safety index rule by using all main body risk values on the supply chain; the blockage point judging and setting module sequentially judges whether the main body risk value of each main body under the supply chain exceeds a preset risk threshold value or not, and sets the corresponding main body as a suspected blockage point when the main body risk value exceeds the preset risk threshold value. The supply chain risk management system can improve the safety evaluation efficiency of the supply chain and can enable managers to make scientific decisions in time according to suspected blockage points.

Description

Supply chain risk management system

Technical Field

The invention relates to a supply chain risk management system.

Background

Enterprises or individuals in the supply chain as main bodies may become a blockage point which affects the normal operation of the whole supply chain due to natural disasters, sudden epidemic situations and other events, at this time, supply chain managers are often required to carry out risk assessment on the whole supply chain, and corresponding measures are made aiming at sparse points (namely, main bodies which have high influence on the blockage point and can dredge the blockage point through a series of measures), so that the supply chain is normally operated.

However, with the rapid development of the economy and the acceleration of the global process in China, China has become the first major country of goods trade in the world. The increasing of cross-border trade is accompanied by the more intricate and complicated composition of the product supply chain, and the increasing speed of the main body involved in the supply chain is increased; with the continuous addition of all parties involved in the supply chain, the blockage and dredging problems involved in the supply chain are diversified, and the blockage and dredging work of the supply chain is harder due to the continuously updated policy and the severe epidemic situation prevention and control task.

Because of the limited administrative resources, the risk assessment of all the subjects involved in the supply chain cannot be performed by the supply chain management personnel; meanwhile, due to the lack of cross-border tracing service, the basis and data involved in the supply chain are not complete enough, so that the responsibility judgment efficiency of the management personnel for the main body is low.

In order to solve the above problems, some simulation systems have appeared at present, which integrate partial blockchain, knowledge graph, and data visualization technologies, but still have many disadvantages, such as:

1) simulation data needs to be manually input or a centralized database is used, so that the data is very easy to lack of authenticity, reliability and public credibility;

2) although simulation models such as prediction and optimization are embedded, the result output by the simulation models is not effectively or reasonably explained, and the influence and the propagation of the influence on the supply chain caused by the implementation simulation cannot be shown, so that the simulation result of the system has low credibility;

3) the risk condition is sensed through a series of indexes, the risk condition can be only roughly judged, and the real-time risk condition judgment cannot be realized;

4) intermediate data are displayed in a traditional chart form, managers cannot obtain deeper data information from the intermediate data, and the managers need to further perform deep data analysis.

In summary, the core functions of the analog simulation system, such as traceability credibility of data, relevance of risk, reliability of analog simulation, and the like, still cannot meet the actual use requirements, and are particularly obvious in the aspect of supply chain safety management, so that managers cannot make scientific decisions quickly to complete supply chain dispersion. Particularly, in an epidemic situation period, the work of dredging the blockage points of the supply chain is very urgent and important, and managers can take scientific and effective decisions in time to relieve the negative effects of the epidemic situation on the whole supply chain to a great extent.

Disclosure of Invention

In order to solve the problems, the invention provides a supply chain risk management system capable of carrying out influence and risk analysis, enforcement simulation and dynamic enforcement recommendation on plugging points and sparse points in a supply chain, which adopts the following technical scheme:

the invention provides a supply chain risk management system, which is used for carrying out risk assessment on a supply chain and each main body under the supply chain so as to enable a supply chain manager to carry out supply chain management based on the risk assessment, and is characterized by comprising the following steps: the supply chain information storage module is used for storing environment information related to a supply chain and all historical traceability information generated by each main body under the supply chain when supply chain business is carried out; the system comprises a map storage module, a knowledge map generation module and a supply chain analysis module, wherein the map storage module is used for storing a knowledge map representing the interrelation among all main bodies in a supply chain; the preprocessing module is used for periodically acquiring historical traceability information and environmental information and preprocessing the historical traceability information and the environmental information to obtain data to be analyzed; the influence module is used for calculating data to be analyzed according to the knowledge graph and a related graph algorithm to obtain the influence of a single main body of each main body and the association degree between the main bodies in the knowledge graph, and calculating the influence degree between each main body and other associated main bodies in sequence according to a preset influence rule to obtain the influence between the main bodies corresponding to all the main bodies; the risk value module is used for analyzing and calculating data to be analyzed according to a preset risk value calculation rule to obtain initial risk values of all the main bodies, and obtaining main body risk values of all the main bodies after risk association according to all the initial risk values and the influence among the main bodies; the safety index module is used for calculating a supply chain safety index of the supply chain corresponding to a plurality of preset safety elements through a preset safety index rule by utilizing all main body risk values on the supply chain; the plugging point judgment setting module is used for sequentially judging whether the main body risk value of each main body under the supply chain exceeds a preset risk threshold value or not and setting the corresponding main body as a suspected plugging point when the main body risk value exceeds the preset risk threshold value; the picture storage module stores a plugging picture; and the display module is used for displaying the blockage point picture and displaying the suspected blockage point, the main risk value corresponding to the suspected blockage point and the supply chain risk value corresponding to the supply chain where the suspected blockage point is located, so that a manager can check the safety state of the supply chain and the condition of the suspected blockage point and manage the supply chain.

The supply chain risk management system provided by the invention can also have the technical characteristics that: the device comprises a main body comprehensive scoring module and a sparse point judging and setting module, wherein the main body comprehensive scoring module utilizes a preset comprehensive scoring rule to carry out comprehensive calculation on the influence of a single main body, the influence among the main bodies and the main body risk value corresponding to each main body in sequence to obtain a main body comprehensive score corresponding to the main body, the sparse point judging and setting module judges whether the comprehensive scores of all the main bodies under a supply chain exceed the preset comprehensive score in sequence and sets the main bodies as suggested sparse points when the comprehensive scores exceed the preset comprehensive score, and the display module displays the suggested sparse points and the influence of the single main body corresponding to the suggested sparse points when displaying a blocked point picture, so that a manager can check the condition of the suggested sparse points and set the supply chain.

The supply chain risk management system provided by the invention can also have the technical characteristics that: the system comprises a history strategy storage module, a main body strategy simulation module and a supply chain strategy simulation module, wherein the history strategy storage module stores history strategy information corresponding to a main body and a history quantitative evaluation table obtained by quantitatively evaluating the history strategy information by using a Driffel method, a picture storage part also stores a strategy selection picture, a display module displays the strategy selection picture to enable a manager to select a suggested sparse point as a strategy object and input corresponding strategy information, the main body strategy simulation module judges whether the strategy information input by the manager belongs to the history strategy information, when the strategy information belongs to the history strategy information, the corresponding blockage point sparse score is obtained according to the history quantitative evaluation table, and when the strategy information does not belong to the history strategy information, the similarity between the strategy information and all the history strategy information is calculated through semantic similarity, and the corresponding blockage point sparse score is obtained in the history quantitative evaluation table according to the history strategy information with the highest similarity The supply chain strategy simulation module carries out risk strategy calculation on the main body risk value based on the plugging point strategy score and the influence among the main bodies of all the corresponding main bodies to obtain the main body strategy score of each main body, integrates all the main body strategy scores to obtain the supply chain strategy score, and the picture storage module also stores a strategy effect picture.

The supply chain risk management system provided by the invention can also have the technical characteristics that the discongesting effect picture also comprises a discongesting knowledge map display part, and the display module displays the discongesting knowledge map picture and displays the discongesting knowledge map display part comprising the knowledge map of the enforcement object, the supply chain score and the main body discongesting score corresponding to each main body, so that managers can check the change generated by the supply chain after enforcement.

The supply chain risk management system provided by the invention can also have the technical characteristics that: the image storage module is used for storing a risk value input image, the display module is used for displaying the risk value input image to enable managers to input expected defibering scores corresponding to the defibering degrees of the supply chains, the decision recommending module is used for analyzing the expected defibering scores by utilizing a pre-trained decision recommending model to obtain suggested break points and decision information corresponding to the expected defibering scores, the image storage module is also used for storing decision recommending images, and once the managers select the expected defibering scores, the display module is used for displaying the decision recommending images and displaying the suggested break points and the decision information to enable the managers to manage the supply chains according to a recommended decision method.

The supply chain risk management system provided by the invention can also have the technical characteristics that the influence rule is as follows:

in the formula of relationship degree weight (1), the relationship degree weight is an influence relationship importance degree value between the subjects, and the calculation of the influence between the subjects comprises the following steps: step A1, calculating data to be analyzed by using a correlation graph algorithm based on the knowledge graph to obtain the influence of a single main body of all main bodies in the knowledge graph and the association degree among all the main bodies; and step A2, sequentially calculating the influence degree between the main body and each other main body in the supply chain containing the main body according to the formula (1) to obtain the risk value between the main bodies corresponding to all the main bodies in the supply chain and each other main body.

The supply chain risk management system provided by the invention can also have the technical characteristics that the risk value calculation rule comprises the following steps: step C1, analyzing the data to be analyzed by a data mining method to obtain a plurality of characteristic parameters of the main body; and step C2, inputting all the characteristic parameters into a machine learning model trained in advance to obtain the initial risk value of each subject.

The supply chain risk management system provided by the invention can also have the technical characteristics that: the safety index rule is used for calculating the safety index corresponding to each safety element according to the main body risk value and the influence weight value of each main body in the supply chain.

The supply chain risk management system provided by the invention can also have the technical characteristics that the preprocessing specifically comprises the following steps: step S1, counting historical traceability information, completing the historical traceability information with less missing items, and deleting the historical traceability information with more missing items; step S2, cross fusion and comparison are carried out on the history tracing information through a machine learning algorithm to obtain cross information; step S3, screening the cross information, modifying or deleting the wrong cross information to obtain the data to be stored; step S4, performing correlation processing on the environment information to obtain correlated environment data; step S5, storing the data to be stored and the data related to the relation analysis in the associated environment data to the map storage module, and storing the data not related to the relation analysis as the data to be analyzed to the traditional relation database, wherein the interrelation between the subjects includes supply chain relation, business relation and conventional relation.

The supply chain risk management system provided by the invention can also have the technical characteristics that the supply chain information storage module is a blockchain which is associated with all main bodies in the supply chain, and the blockchain is used for automatically acquiring the related filing information, the authentication information and the supply chain related basic information of the main bodies and the supply chain related public opinion information and the environment public opinion information on each website, and correspondingly storing the related filing information, the authentication information and the supply chain related basic information as history traceability information and the supply chain related public opinion information and the environment public opinion information as environment information.

Action and Effect of the invention

According to the supply chain risk management system, the data to be analyzed can be analyzed and evaluated to obtain the supply chain risk value and the suspected blockage point, so that the working efficiency of supply chain safety evaluation is improved, and managers can make scientific decisions according to the suspected blockage point and the environmental information to further complete supply chain dredging work, particularly the supply chain dredging work under epidemic situations.

The influence module calculates the influence of a single main body of each main body in the knowledge graph and the association degree between the main bodies according to the knowledge graph and the data to be analyzed through a related graph algorithm, and calculates the influence degree between each main body and other associated main bodies in sequence through a preset influence rule to obtain the influence between the main bodies corresponding to all the main bodies, so that managers can know the mutual influence condition between the main bodies.

And the risk value module analyzes and calculates the data to be analyzed according to a preset risk value calculation rule to obtain initial risk values of all the main bodies, and obtains the main body risk values of all the main bodies after risk association according to all the initial risk values and the influence among the main bodies, so that managers can know the risk state of all the main bodies.

And the safety index module calculates the supply chain safety indexes corresponding to a plurality of preset safety elements of the supply chain by utilizing all main body risk values on the supply chain through a preset safety index rule, so that a multi-dimensional, deep-level and penetrating correlation analysis result is obtained, and further, a manager can know the whole safety state of the supply chain through the supply chain risk values.

And then, the plugging point judgment and setting module sequentially judges whether the risk value of each main body exceeds a preset risk threshold value or not, and sets the corresponding main body as a suspected plugging point when the risk value of each main body exceeds the preset risk threshold value, so that a manager can quickly lock the suspected plugging point and take measures for the suspected plugging point.

Further, the supply chain information storage module is a block chain associated with all main bodies in the supply chain, and the environmental information and the historical traceability information can be automatically acquired from the block chain, so that manual input by a manager is not needed, and the authenticity, reliability and public credibility of the information can be ensured.

And finally, the sparse knowledge map picture is displayed, so that the manager can clearly master the deep information such as the risk source, the enforcement propagation path, the result and the like of the supply chain without further deep data analysis by the manager.

Drawings

Fig. 1 is a block diagram of a supply chain risk management system according to a first embodiment of the present invention;

FIG. 2 is a flowchart illustrating a process of calculating an influence between subjects according to a first embodiment of the present invention;

FIG. 3 is a diagram illustrating a quantitative evaluation rule of a discongesting method according to a first embodiment of the invention;

FIG. 4 is a flowchart of a process for determining a cut-off untwining score according to an embodiment of the present invention

Fig. 5 is a flowchart of a supply chain risk management system according to a first embodiment of the present invention; and

fig. 6 is a flowchart of a supply chain risk management system according to a second embodiment of the present invention.

Detailed Description

In order to make the technical means, the inventive features, the objectives and the functions of the present invention easy to understand, the supply chain risk management system of the present invention is specifically described below with reference to the embodiments and the accompanying drawings.

< example one >

Fig. 1 is a block diagram of a supply chain risk management system according to an embodiment of the present invention.

As shown in fig. 1, the supply chain risk management system 100 includes a control module 10, a supply chain information storage module 11, a map storage module 12, a preprocessing module 13, an influence module 14, a risk value module 15, a safety index module 16, a blockage point judgment setting module 17, a main body comprehensive scoring module 18, a sparse point judgment setting module 19, a historical strategy storage module 20, a main body strategy simulation module 21, a supply chain strategy simulation module 22, a screen storage module 23, a display module 24, and a weight storage module 25.

The supply chain risk management system 100 controls the following modules via the control module 10.

The supply chain information storage module 11 stores environment information related to a supply chain and historical traceability information related to supply chain services of each main body under the supply chain.

The supply chain information storage module 11 is a blockchain associated with all the entities in the supply chain, and the blockchain is used for automatically acquiring the relevant filing information, the authentication information, the relevant basic information of the supply chain and the relevant public opinion information and the environmental public opinion information of the supply chain on each website, and correspondingly storing the relevant filing information, the authentication information and the relevant basic information of the supply chain as history traceability information and the relevant public opinion information and the environmental public opinion information of the supply chain as environmental information.

Wherein, the main body refers to enterprises, products, personnel and the like involved in the supply chain; the environment information comprises real-time public sentiment and epidemic information, policy information, laws and regulations of enterprises, products, personnel and other main bodies related to the whole supply chain; the historical traceability information comprises related filing information and authentication information on the trusted block chain, and related basic information of enterprises, products, personnel and the like related to the whole supply chain; the system further comprises information of the trusted cloud system and information on the trusted cloud block chain, wherein enterprises needing to be accessed by the trusted cloud system record on the trusted cloud system, and upload traceability information in the circulation process of the supply chain for products so as to form main body traceability information corresponding to the main body.

In this embodiment, the environmental information is obtained from a government website, an enterprise website, and a third-party website through a web crawler technology; historical traceability information is obtained from a trusted cloud blockchain (i.e., the blockchain) and a trusted cloud system.

The profile storage module 12 stores knowledge profiles representing the interrelationships between various entities in the supply chain.

The knowledge graph is a visual display of the interconnection between the various entities in each supply chain generated according to the data to be analyzed, so that a non-professional technician can know the relationship (such as a cooperative relationship, an employment relationship and the like) between the various entities in the whole supply chain.

The interrelationship between the main bodies comprises a supply chain relation, a business relation and a conventional relation.

In this embodiment, the supply chain relationship is directed to a relationship between an enterprise and an enterprise, specifically, an enterprise and an upstream enterprise, a downstream enterprise cooperating with the enterprise, and a business-like enterprise having a business-like relationship with the enterprise.

The business relationship is aimed at the relationship between products and enterprises, enterprises and risks, personnel and risks, enterprises and public opinions and personnel and public opinions. Specifically, a product and a supplier company, a producer company, a distributor company, a retailer company, a logistics company, a risk and a company or person related to the risk, a public opinion and a company or person related to the public opinion, and the like of the product.

Conventional relationships are directed to relationships between businesses and enterprises, personnel and businesses, and in particular, businesses and collaborating enterprises with the business, businesses and other enterprises in which the business holds shares, personnel and businesses in which the personnel has a job, the personnel is on the job director, manager, high manager, etc. of the business in which the personnel has a job.

The preprocessing module 13 may periodically obtain the historical traceability information and the environmental information and perform preprocessing to obtain the data to be analyzed.

The pretreatment specifically comprises the following steps:

and step S1, counting the history tracing information, completing the history tracing information with less missing items, and deleting the history tracing information with more missing items.

In this embodiment, when the history traceability information missing item is not more than 3 items, the preprocessing module performs completion automatically according to the service logic or performs completion by using the similar history traceability information nearby, for example, when a product lacks a production date, the production date of other products in the same batch is used as the production date of the product; when the history tracing information missing items are more than 3, the history tracing information is directly deleted, so that the accuracy of the history tracing information is ensured.

And step S2, cross fusion and comparison are carried out on the history tracing information through a machine learning algorithm to obtain cross information.

In this embodiment, a K-means clustering algorithm is used for fusing data among classes for multi-source historical traceability information.

And step S3, screening the cross information, and modifying or deleting the wrong cross information to obtain the data to be stored.

In this embodiment, when the cross information is inconsistent, the preprocessing module 13 automatically selects the information obtained from the authoritative channel, and discards the information from the channel with relatively low authority. When a legal value is judged, the legal value can be determined by setting the legal time of the date.

In step S4, the environment information is correlated to obtain correlated environment data.

In this embodiment, association processing is performed on enterprises, products, personnel and the like which can reach in the environment information to obtain associated environment data, for example, association processing is required for enterprises, products, personnel and the like which are influenced by epidemic situations.

Step S5, store the data to be stored and the data related to the relational analysis in the associated environment data to the atlas storage module 12, and store the data not related to the relational analysis as the data to be analyzed to the traditional relational database (e.g., MySQL, oracle).

In this embodiment, a knowledge map is formed by the steps of information extraction, knowledge fusion, knowledge processing, and the like for data relating to the relational analysis, and is stored in the map database Neo4J (i.e., the map storage module 12).

In this embodiment, the preprocessing further includes adding new data, correcting erroneous data, completing missing data, discarding useless data, and sorting required service features. For example, the price safety index of the supply chain is arranged by extracting the prices of key links or important commodities in the whole supply chain, establishing a price index model, and viewing the safety condition of the whole supply chain from the perspective of the change condition of the market price of the important commodities in the whole supply chain; for the content related to the relationship analysis, the main bodies (enterprises, products, personnel) in the supply chain, the relationship (cooperative relationship, employer relationship, etc.) between the main bodies are extracted correspondingly.

The influence module 14 calculates the influence of a single main body of each main body and the association degree between the main bodies in the knowledge map according to the knowledge map and the data to be analyzed by a correlation map algorithm, and calculates the influence degree between each main body and each other associated main body in sequence according to a predetermined influence rule to obtain the influence between the main bodies corresponding to all the main bodies.

The influence module 14 deeply mines the relationship between the subjects of the knowledge graph through a correlation graph algorithm, and deeply analyzes the influence of each subject, the risk value of the subject, the tightness between the subjects, the risk transfer between the subjects, and the like under each supply chain. For example, the strength of the relationship between the subjects is determined on the knowledge graph by the PageRank and PersonRank algorithms.

Fig. 2 is a flowchart of a process of calculating an influence between subjects according to a first embodiment of the present invention.

As shown in fig. 2, the calculation of the influence between the subjects includes the following steps:

and A1, calculating the data to be analyzed by using a correlation diagram algorithm based on the knowledge diagram to obtain the influence of a single main body of all main bodies in the knowledge diagram and the association degree among the main bodies.

And A2, sequentially calculating the influence among the main bodies between the main body and the other N-1 main bodies according to the influence rule formula (1) to obtain all the influence among the main bodies of the N main bodies, wherein each main body corresponds to the influence among the N-1 main bodies.

Influence between subjects (single subject influence degree relation degree weight (1)

The relationship degree weight is an importance degree value of the influence relationship among the subjects, and can be divided into different levels such as degree weight resetting of the first degree relationship, degree weight weighting of the second degree relationship and the like.

In this embodiment, assuming that the influence of the single body a is 50, the degree of association between the single body a and the single body B is 0.8, the degree of association between the single body B and the single body C is 0.7, the degree-of-one-degree weight is 0.9, and the degree-of-two-degree weight is 0.6, the influence of the single body a on the single body B is 50 × 0.8 × 0.9 — 36, and the influence of the single body a on the single body C is 50 × 0.9 × 0.7 — 0.6 — 16.

The risk value module 15 analyzes and calculates the data to be analyzed according to a predetermined risk value calculation rule to obtain initial risk values of each subject, and obtains subject risk values of each subject after risk association according to all the initial risk values and the inter-subject influence among the subjects.

Wherein the risk value calculation rule comprises the following steps:

and step C1, analyzing the data to be analyzed by using a data mining method to obtain a plurality of characteristic parameters of the main body.

In this embodiment, the data mining method includes a plurality of data mining methods such as statistical analysis and cluster analysis.

And step C2, inputting all the characteristic parameters into a machine learning model trained in advance to obtain the initial risk value of each subject.

In this embodiment, the history tracing information is analyzed by a data mining method to obtain a series of characteristic parameters (e.g., enterprise registered funds, business state, personnel scale, business change condition, yield change, deviation, high violation group member, etc.) such as the history supply and demand condition and risk change condition of the main body. And performing model training by taking the characteristic parameters as a training set of the machine learning GBDT model until convergence to obtain a trained machine learning model.

The data mining method at least comprises statistical analysis and cluster analysis.

The statistical analysis is to analyze the data to be analyzed to obtain the overall business information related to the supply chain business (such as the supply chain general situation, the product flow direction information and the enterprise participation information, the country of origin of the supply chain product, the production enterprise, the transportation enterprise and the review unit).

The clustering analysis is to perform clustering analysis on the whole business information by using a K-means clustering algorithm, divide all subjects into different risk categories (for example, divide production enterprises into three risk categories of high/medium/low), and perform key supervision on enterprises in a production enterprise group divided into high risks.

In this embodiment, the subject risk value between each associated subject may change as a whole due to the change of the subject risk value of a certain subject, such as: assuming that the initial risk value of the subject X1 is 40, the initial risk value of the subject X2 is 33, and the degree of association between X1 and X2 is 0.7 as calculated by the influence module 14, when there is a bad public opinion occurrence that directly gives the subject X1 the risk value of 20, the subject risk value of the subject X1 becomes 40+ 20-60, and the subject risk value of the subject X2 associated with the subject X1 becomes 33+ 0.7-20-47.

The safety index module 16 calculates a supply chain safety index corresponding to a predetermined plurality of safety elements in the supply chain according to a predetermined safety index rule by using all subject risk values in the supply chain.

The safety elements need to be preset by management personnel according to actual requirements, and include production, funds, trade, logistics, sanitation and the like related to a supply chain.

In the embodiment, the risk scores of different raw materials are obtained through the GBDT algorithm, and the safety condition of the production end is determined through rules such as weighting according to the importance degree of the different raw materials in the supply chain.

The weight storage module 25 stores a plurality of influence weight values corresponding to each security element and corresponding to the degree of importance of the risk involved in each subject in the security element.

Wherein the influence weight value is preset by a manager according to the risk importance degree of all the main bodies related to the safety elements in different safety elements,

the safety index rule is used for calculating the safety index corresponding to each safety element according to the main body risk value and the influence weight value of each main body in the supply chain.

In this embodiment, the safety index rules are different in calculating safety indexes of different safety elements, for example, when calculating a production safety index for production of a safety element of a supply chain, the safety index rules pay attention to raw material risks, production enterprise risks, employee risks, and the like, and a production safety index is calculated according to the contribution degree of each risk to the production safety index. And in the process of calculating the safety indexes of other safety elements, carrying out weighted calculation according to the importance degrees of different risks involved by each safety element to obtain the safety index corresponding to each safety element.

In this embodiment, taking the calculation of the production safety index of the safety element production of the supply chain as an example: production safety index ═ raw material 1 risk (e.g., the most major raw material for making steel) × 0.4+ raw material 2 risk (e.g., the minor raw material for making steel) × 0.3+ manufacturing enterprise risk (e.g., steel manufacturing enterprise risk value) × 0.2+ employee risk × 0.1.

Wherein, the raw material 1 risk, the raw material 2 risk and the production enterprise risk are obtained by using a data mining model or a machine learning model. Taking a machine learning model as an example: if the supply chain has historical data of the raw material 1, the risk condition of the raw material 1 is learned through the machine learning model, and when the latest data of the raw material 1 on the supply chain is input into the trained machine learning model, the risk condition of the raw material 1 at the moment is obtained. Wherein, the calculation formula involved is as follows:

in the formula, S is a risk value of the raw material 1, A is the sum of the current-stage supply quantity, namely the import quantity of the raw material 1 at the current stage and the domestic yield of the raw material 1 at the current stage, B is the sum of the current-stage supply quantity, namely the import quantity of the raw material 1 at the same stage and the domestic yield of the raw material 1 at the same stage, and the time period can be selected from week or month and the like according to the situation.

Raw material 1 risk value has a normal interval: l ≦ S ≦ M, where M ═ the contemporaneous supply amount B ≦ the annual maximum deviation, L ═ the contemporaneous supply amount B ≦ the annual minimum deviation, the annual maximum deviation (BIAS) ═ MAX (monthly supply/yearly average supply 1 to 12), and the annual minimum deviation ≦ Min (monthly supply/yearly average supply 1 to 12).

In this embodiment, when the calculated risk value of the raw material 1 is 65, it is determined that the raw material 1 is at a low risk (the low risk category is determined according to the upper limit < S ═ 1.1 or the lower limit < 0.9 ═ S < lower limit); judging that the raw material 1 is at an intermediate risk when the raw material 1 risk value is 75 (the intermediate risk category is according to an upper limit of 1.1< S ═ upper limit of 1.2 or a lower limit of 0.8 ═ S < lower limit of 0.9); raw material 1 is judged to be at high risk when raw material 1 risk value is 90 (high risk category according to upper limit 1.2< S or S < lower limit 0.8).

The plugging point judgment setting module 17 sequentially judges whether the main body risk value of each main body under the supply chain exceeds a preset risk threshold value, and sets the corresponding main body as a suspected plugging point when the main body risk value exceeds the preset risk threshold value.

The subject comprehensive scoring module 18 sequentially performs comprehensive calculation on the individual subject influence, the inter-subject influence and the subject risk value corresponding to each subject by using a predetermined comprehensive scoring rule (for example, a weighted average) to obtain a subject comprehensive score corresponding to the subject.

The sparse point judgment setting module 19 sequentially judges whether the comprehensive scores of all the subjects in the supply chain exceed the preset comprehensive score or not, and sets the subjects as the suggested sparse points when the comprehensive scores exceed the preset comprehensive score.

In this embodiment, when the subject composite score exceeds the predetermined composite score, the risk value of the subject and the influence of the single subject are both in a high state.

Fig. 3 is a schematic diagram of a quantitative evaluation rule of a discongesting method according to a first embodiment of the invention.

The history policy storage module 20 stores history policy information corresponding to the body and a history quantization score table obtained by quantitatively scoring the history policy information by the dfield method.

The historical strategy information comprises a strategy tool and a dismissal method. The enforcement tool is an enforcement policy aiming at common blockage problems in the supply chain, for example, when the supply chain problem is a fund problem, the enforcement tool subsidies the fund; when the supply chain problem is a tax issue, the enforcement tool is a tax offer (as shown in FIG. 3).

The enforcement tools include exemption of house rents, tax benefits, social security, special loans, fund subsidies and the like. For example, the relief method corresponding to the exemption of the tenants includes direct exemption, direct subsidy, return by ratio, and the like; the corresponding untwining methods of the tax preferential include tax deduction, return according to proportion, delayed payment and the like; the corresponding relief method of the social security comprises direct subsidy, delayed payment and the like; the corresponding relief method of the special loan comprises interest loan, low interest loan and the like; the fund subsidy corresponds to a mediation method such as direct subsidy and indirect subsidy.

In this embodiment, the historical quantitative rating table is obtained by quantifying according to the experience summary of the expert on the historical applied information by using the dfield method, and each item of historical applied information in the historical quantitative rating table corresponds to one blockage evacuation score (as shown in fig. 3).

The main body policy making simulation module 21 determines whether the policy information input by the administrator belongs to the historical policy information, obtains a corresponding blockage point untwining score according to the historical quantitative scoring table when the policy information belongs to the historical policy information, calculates the similarity between the policy information and all the historical policy information according to the semantic similarity when the policy information does not belong to the historical policy information, and obtains a corresponding blockage point untwining score in the historical quantitative scoring table according to the historical policy information with the highest similarity.

Fig. 4 is a flowchart of a process of determining a blockage untwining score according to a first embodiment of the present invention.

As shown in fig. 4, the process of determining the obstruction point untwining score includes the following steps:

step B1, judging whether the administrator input strategy information belongs to the historical strategy information, if so, entering step B3, and if not, entering step B2;

step B2, calculating the similarity between the construction information and all the historical construction information by using a semantic similarity method to obtain a plurality of similarity values, selecting the historical construction information with the highest similarity value as a basis for searching blockage evacuation scores in a historical quantitative score table of the construction information, and then entering step B3;

and step B3, calculating a blockage point untwining score corresponding to the historical imposing information in a historical quantitative scoring table according to the historical imposing information.

The supply chain enforcement simulation module 22 performs risk relief calculation on the subject risk values based on the blockage relief scores and the inter-subject influence of all the corresponding subjects to obtain the subject relief scores of each subject, and integrates all the subject relief scores to obtain the supply chain relief scores.

In this embodiment, the influence degrees between the subjects are obtained by using a correlation graph algorithm, and a regulatory calculation process is performed in combination with the business features, for example, after the influence degrees between the subjects are obtained, a supply chain business rule — that "the influence of a large enterprise on a small enterprise in a supply chain is greater than the influence of a small enterprise on the large enterprise", is performed, and a corresponding weighting process is performed in combination with the influence degrees between the nodes in combination with the importance degrees of the nodes.

The supply chain planning simulation module 22 performs planning simulation on the suspected blockage points according to the blockage point untwining scores and the influence degrees among all the main bodies obtained by using the correlation graph algorithm, calculates and obtains main body untwining scores corresponding to all other main bodies in the supply chain where the suspected blockage points are located, and integrates all the main body untwining scores to obtain the supply chain untwining scores. For example, the initial subject risk value of the suspected blockage point X3 is 64, the individual subject influence is 80, the initial subject risk value of the subject X4 associated with the suspected blockage point X3 is 61, the degree of association between the X3 and the X4 is 0.8, and the blockage point untwining score corresponding to the selected untwining method (for example, if the enforcement tool is selected as a tax subsidy, and the untwining method is selected as a direct subsidy, the score corresponding to the untwining method is-30, that is, the risk value of the enforcement object can be reduced by 30 points by implementing the tax subsidy on the enforcement object in a direct subsidy manner) is 30, then the blockage point X3 becomes a normal subject X3 and the corresponding subject risk value thereof is 64-33-34, while the blockage point solution score of the subject X4 is 30-80% × 0.8-19.2, and the subject risk value of the subject X4 becomes 61-19.8-41.2.

In this embodiment, the supply chain policy simulation module 22 sets the supply chain untwining score and the corresponding untwining method, the policy tool corresponding to the untwining method, and the policy object corresponding to the policy tool as a set of historical policy information to be stored in the historical policy storage module 20.

In this embodiment, the supply chain strategy simulation module 22 further summarizes the strategy according to the supply chain discongest score, for example, the range of the strategy influence, the evaluation of the discongesting effect, and the like.

The frame storage module 23 stores a block point frame, a block point knowledge map frame, a strategy selection frame, a dismissal effect frame, and a dismissal knowledge map frame.

The blockage point screen is used for displaying a suspected blockage point, a main body risk value corresponding to the suspected blockage point, a supply chain risk value corresponding to a supply chain where the suspected blockage point is located, a suggested sparse point and a single main body influence corresponding to the suggested sparse point.

In this embodiment, the block point screen displays a supply chain risk value, a safety index of each safety element of the supply chain, a subject name corresponding to the suspected block point, a subject risk value corresponding to the suspected block point, a subject name corresponding to the suggested sparse point, a single subject influence corresponding to the suggested sparse point, subject names of all subjects related to the suspected block point, an influence condition influenced by the suspected block point, a relationship attribute (such as direct correlation, indirect correlation, and the like) with the suspected block point, and subject details (such as type, address, and the like) corresponding to the suspected block point.

The strategy selection screen allows a manager to select a suggested sparse point as a strategy object and input strategy information.

The strategy information includes a plurality of strategy tools corresponding to each strategy object and a plurality of defibering methods corresponding to each strategy tool.

The untwining effect picture is used for displaying the supply chain untwining score and the untwining knowledge map display part.

In this embodiment, the untwining effect screen displays the supply chain untwining score, the summary of the entire supply chain untwining (e.g., the range of the influence of the policy, the evaluation of the untwining effect, etc.), and the benefit (e.g., the reduced risk value) of each associated subject from the supply chain untwining score.

The untwining knowledge graph display part is used for displaying a knowledge graph containing the strategy objects, the supply chain scores, the subject names corresponding to the subjects and the subject untwining scores corresponding to the subjects.

The display module 24 is used for the above pictures, thereby completing human-computer interaction.

Fig. 5 is a flowchart of a supply chain risk management system according to a first embodiment of the present invention.

As shown in fig. 5, the supply chain risk management system includes the following steps:

step T1, the supply chain information storage module 11 periodically obtains the historical traceability information and the environmental information, and pre-processes the information by the pre-processing module 13 to obtain the data to be analyzed, and then the step T2 is performed;

step T2, the influence module 14 calculates and analyzes the data to be analyzed by using the knowledge graph, the related graph algorithm and the influence rule stored in the graph storage module 12 to obtain the influence of a single main body and the influence among the main bodies, and then the step T3 is performed;

step T3, the risk value module 15 analyzes and calculates the data to be analyzed by using the risk value calculation rule to obtain the initial risk value of each subject, and obtains the subject risk value according to the influence between subjects, and then the step T4 is performed;

step T4, the safety index module 16 calculates a supply chain safety index of each supply chain safety element corresponding to the supply chain according to a predetermined safety index rule by using all the subject risk values on the supply chain, and then proceeds to step T5;

step T5, the block point judgment setting module 17 determines whether the risk value of the subject exceeds a predetermined risk threshold, if so, sets the subject as a suspected block point and proceeds to step T6, and if not, repeats step T5 and determines the next subject;

step T6, the subject comprehensive scoring module 18 scores the subject information by using a comprehensive scoring rule based on the subject influence and the subject risk value to obtain a subject comprehensive score, and then the step T7 is performed;

step T7, the sparse point judgment setting module 19 judges whether the subject comprehensive score is larger than the preset comprehensive score, if yes, the subject is set as the suggested sparse point and enters step T8, and if not, the step T7 is repeated and the next subject is judged;

step T8, selecting the suggested sparse points as the construction objects in the main body construction simulation module 21, inputting construction information to perform main body construction simulation to obtain the blockage sparse scores, and then entering step T9;

step T9, the supply chain enforcement simulation module 22 performs scoring by using a correlation graph algorithm based on the blockage evacuation score, the subject risk value, and the subject influence to obtain a supply chain evacuation score, and then proceeds to step T10;

at step T10, the display module 24 displays the suspected blockage point, the proposed evacuation point, the supply chain risk value, and the supply chain evacuation effect, and then enters an end state.

< example two >

In the first embodiment, in order to shorten the time for the administrator to select the discourse method, the history policy storage module, the main body policy simulation module, and the supply chain policy simulation module in the first embodiment directly obtain information such as the discourse method, the policy tool, the policy object, and the like corresponding to the desired discourse score according to the desired discourse score that the administrator wants the supply chain to reach, and directly implement policy, so as to complete the policy simulation of the present invention.

For convenience of expression, in the second embodiment, the same reference numerals are given to the same structures as those in the first embodiment, and the same descriptions are omitted.

The strategy recommending module analyzes the expected untwining score by using a pre-trained strategy recommending model to obtain a suggested sparse point and strategy information corresponding to the expected untwining score.

The strategy recommendation model is trained through a machine learning algorithm based on historical strategy information, and can acquire the corresponding relation between the expected fluffing score and the strategy information and between the expected fluffing score and the suggested fluffing points.

The application information includes multiple grooming methods, different application tools corresponding to the multiple grooming methods, and different suggested grooming points corresponding to the multiple application tools.

In this embodiment, the enforcement recommendation model obtains the association relationship among the suggested sparse points, the enforcement tool, the grooming method, and the expected grooming scores through an Aprior algorithm.

The picture storage module is also used for storing a risk value input picture and a strategy recommendation picture.

The risk value input screen is used for the manager to input a desired untwining score corresponding to the untwined degree of the supply chain.

The strategy recommending picture is used for displaying the suggested sparse points and the strategy information.

The display module is used for displaying the pictures to realize human-computer interaction.

As shown in fig. 6, the supply chain risk management system includes the following steps:

step E1, the supply chain information storage module periodically acquires historical traceability information and environmental information, and preprocesses the historical traceability information and the environmental information through the preprocessing module to obtain data to be analyzed, and then the step E2 is carried out;

e2, the influence module calculates and analyzes the data to be analyzed by using the knowledge graph, the related graph algorithm and the influence rule in the graph storage module to obtain the influence of a single main body and the influence among the main bodies, and then the step E3 is carried out;

e3, analyzing and calculating the data to be analyzed by the risk value module by using a risk value calculation rule to obtain an initial risk value of each main body, obtaining a main body risk value according to influence among the main bodies, and entering the step E4;

step E4, the safety index module calculates the supply chain safety index of each supply chain safety element corresponding to the supply chain by using all the subject risk values on the supply chain through a preset safety index rule, and then the step E5 is carried out;

step E5, the block point judgment setting module judges whether the risk value of the main body exceeds the preset risk threshold value, if yes, the main body is set as a suspected block point and the step E6 is carried out, and if not, the step E5 is repeated to judge the next main body;

e6, the subject comprehensive scoring module scores the subject by using a comprehensive scoring rule based on the subject influence and the subject risk value to obtain a subject comprehensive score, and then the step E7 is carried out;

e7, judging whether the main body comprehensive score is larger than the preset comprehensive score by the sparse point judgment setting module, if so, setting the main body as a suggested sparse point and entering the step E8, and if not, repeating the step E7 to judge the next main body;

step E8, inputting the desired fluffing score on the risk value input picture, and then entering step E9;

step E9, the enforcement recommendation module analyzes a suggestion sparse point corresponding to the expected fluffing score and enforcement information corresponding to the suggestion sparse point by utilizing an enforcement recommendation model based on the expected fluffing score to serve as an enforcement recommendation result, and then the step E10 is carried out;

step E10, the display module displays the suspected blockage situation, the supply chain safety status and the enforcement recommendation result, and then enters the end status.

Examples effects and effects

According to the supply chain risk management system provided by the embodiment, the influence module calculates the influence of a single main body of each main body and the association degree between the main bodies according to the knowledge graph and the data to be analyzed through a related graph algorithm, and calculates the influence degree between each main body and other associated main bodies in sequence through the preset influence rule to obtain the influence between the main bodies corresponding to all the main bodies, so that the management personnel can know the mutual influence condition between the main bodies.

And then the blockage point judging and setting module sequentially judges whether the main body risk value of each main body under the supply chain exceeds a preset risk threshold value or not, and sets the corresponding main body as a suspected blockage point when the main body risk value exceeds the preset risk threshold value, so that a manager can quickly lock the suspected blockage point and take measures for the suspected blockage point.

In addition, in the embodiment, the supply chain information storage module is a blockchain associated with all the main bodies in the supply chain, and the blockchain automatically acquires information, so that the historical traceability information and the environmental information can be continuously updated, and the data and the basis have higher reliability and better integrity.

In addition, in the embodiment, because the suspected blockage untwining score and the supply chain untwining score are obtained through the main body enforcement simulation module and the supply chain enforcement simulation module by inputting the enforcement object selected by the manager and the enforcement information, the manager performs the analog simulation of enforcement through the system, and sees the expected effect generated after enforcement in advance, so that the manager can determine which untwining method to adopt according to the expected effect to realize scientific and accurate management of the supply chain.

In addition, in the embodiment, the implementation recommending module outputs the untwining method corresponding to the expected untwining score, the implementation tool corresponding to the untwining method and the suggested untwining point corresponding to the implementation tool according to the expected untwining score input by the manager, so that the manager can directly obtain the corresponding untwining method from the desired implementation effect through the system, the time for analyzing the supply chain and judging which untwining method is adopted is greatly saved, and the problem of supply chain blockage is solved in time.

The above-described embodiments are merely illustrative of specific embodiments of the present invention, and the present invention is not limited to the description of the above-described embodiments.

In the above embodiment, the screening of the cross data in the preprocessing module is the automatic screening of the preprocessing module, and the invention can also be manually screened by a manager so as to ensure that the data to be stored has higher accuracy.

In the above embodiment, the suggested sparse point is obtained by the sparse point judgment setting module, and the administrator can select another main body belonging to the same supply chain as the suspected blockage point as the suggested sparse point, so as to perform strategy simulation on the suggested sparse point.

In the above embodiment, the display module displays a plurality of screens for the administrator to view, select, input, and the like, but the present invention is not limited to the above screens, and for example, information such as suspected congestion points, suggested sparse points, strategy simulation, strategy recommendation, and the like may be displayed in the same screen.

In the above embodiment, the influence rule is used for weighting the degree of relationship between the body influence and the single body influence, and the formula may be adjusted according to the actual situation to obtain the body influence more suitable for the actual application.

In the above embodiment, data is stored in advance through a block chain, and in other schemes, the present invention may also directly access various data sources and results of data analysis (for example, data in a supply chain business scenario and a customs cross-border business scenario), so as to directly perform a main risk analysis by using the data (for example, a main risk analysis at a bottom layer may access a packet listening applet for analysis, and a main risk value of the packet listening analysis is directly used).

In the above embodiment, the risk value calculation rule is to analyze the data to be analyzed by a data mining method and then learn by a machine learning model to obtain an initial risk value, and the method of the present invention may also use the result obtained by the data mining analysis as the input of the supply chain business rule and the machine learning model, and finally weight the result of the supply chain business rule and the result of the machine learning model to obtain a final initial risk value.

In the above embodiment, the employee risk is considered in the calculation of the production safety index, and the production safety index of only one enterprise is calculated, but in other schemes, the production safety index calculation can be performed by omitting the employee risk, and the final production safety index is obtained by performing weighted calculation on the production safety indexes of a plurality of enterprises by using the supply chain business rules.

Claims

1. A supply chain risk management system for performing risk assessment on a supply chain and various entities under the supply chain to allow supply chain management by supply chain management personnel based on the risk assessment, comprising:

the supply chain information storage module is used for storing environment information related to the supply chain and all historical traceability information generated by each main body under the supply chain when supply chain business is carried out;

the map storage module is used for storing a knowledge map representing the interrelation among the main bodies in the supply chain;

the preprocessing module is used for periodically acquiring the historical traceability information and the environmental information and preprocessing the historical traceability information and the environmental information to obtain data to be analyzed;

the influence module is used for calculating the data to be analyzed according to the knowledge graph and a related graph algorithm to obtain the influence of a single main body of each main body and the association degree between the main bodies in the knowledge graph, and sequentially calculating the influence degree between each main body and other associated main bodies according to a preset influence rule to obtain the influence between the main bodies corresponding to all the main bodies;

the risk value module is used for analyzing and calculating the data to be analyzed according to a preset risk value calculation rule to obtain an initial risk value of each main body, and obtaining a main body risk value of each main body after risk association according to all the initial risk values and the influence among the main bodies;

the safety index module is used for calculating a supply chain safety index of the supply chain corresponding to a plurality of preset safety elements by utilizing all the main risk values on the supply chain through a preset safety index rule;

the plugging point judgment setting module is used for sequentially judging whether the main body risk value of each main body under the supply chain exceeds a preset risk threshold value or not and setting the corresponding main body as a suspected plugging point when the main body risk value exceeds the preset risk threshold value;

the picture storage module stores a plugging picture; and

and the display module is used for displaying the blockage point picture and displaying the suspected blockage point, the main risk value corresponding to the suspected blockage point and the supply chain risk value corresponding to the supply chain where the suspected blockage point is located, so that the administrator can check the safety state of the supply chain and the condition of the suspected blockage point and manage the supply chain.

2. The supply chain risk management system of claim 1, further comprising:

a main body comprehensive scoring module and a sparse point judgment setting module,

wherein the main body comprehensive scoring module sequentially carries out comprehensive calculation on the single main body influence, the inter-main body influence and the main body risk value corresponding to each main body by utilizing a preset comprehensive scoring rule to obtain a main body comprehensive score corresponding to the main body,

the sparse point judgment setting module sequentially judges whether the comprehensive scores of all the main bodies in the supply chain exceed the preset comprehensive score or not and sets the main bodies as suggested sparse points when the comprehensive scores exceed the preset comprehensive score,

when the display module displays the plugging point picture, the suggested sparse points and the single main body influence force corresponding to the suggested sparse points are also displayed, so that the manager can check the suggested sparse points and can make a policy on the supply chain.

3. The supply chain risk management system of claim 2, further comprising:

a history execution storage module, a main execution simulation module and a supply chain execution simulation module,

wherein the history strategy storage module stores history strategy information corresponding to the subject and a history quantitative score table obtained by quantitatively scoring the history strategy information by a Defield method,

the frame storage part also stores the selected frame,

the display module displays the strategy selection picture to enable the manager to select one suggested sparse point as a strategy object and input corresponding strategy information,

the main body policy applying simulation module judges whether the policy applying information input by the manager belongs to the historical policy applying information,

when the enforcement information belongs to the historical enforcement information, obtaining a corresponding blockage untwining score according to the historical quantitative scoring table,

when the enforcement information does not belong to the historical enforcement information, calculating the similarity between the enforcement information and all the historical enforcement information through semantic similarity, and obtaining the corresponding blockage evacuation score in the historical quantitative scoring table according to the historical enforcement information with the highest similarity,

the supply chain strategy simulation module carries out risk solution calculation on the subject risk value based on the blockage point solution scores and the influence among the subjects corresponding to all the subjects so as to obtain the subject solution scores of all the subjects, and integrates all the subject solution scores to obtain the supply chain solution scores,

the picture storage module also stores a fluffing effect picture,

once the manager selects the suggested dismissal point and inputs the curation information, the display module displays a dismissal effect screen and displays the supply chain dismissal score and the body dismissal score of each of the bodies, so that the manager can predict the effect according to the selected curation object and the input curation information.

4. The supply chain risk management system of claim 3, wherein:

wherein the dismissal effect picture also comprises a dismissal knowledge map display part,

the display module displays the picture of the defibering knowledge graph and displays the display part of the defibering knowledge graph containing the knowledge graph of the planning object, the supply chain scores and the main body defibering scores corresponding to the main bodies, so that the manager can view changes generated by the supply chain after planning.

5. The supply chain risk management system of claim 2, further comprising:

a policy-applying recommending module for recommending the policy-applying module,

wherein the picture storage module is also used for storing a risk value input picture,

the display module displays the risk value input screen for the manager to input a desired untwining score corresponding to a degree of untwining of the supply chain,

the strategy recommending module analyzes the expected untwining score by utilizing a pre-trained strategy recommending model to obtain the suggested sparse points corresponding to the expected untwining score and the strategy information,

the picture storage module also stores a enforcement recommendation picture,

once the manager selects the desired untwining score, the display module displays the curation recommendation screen and displays the suggested untwining points and the curation information, so that the manager manages the supply chain according to a recommended curation method.

6. The supply chain risk management system of claim 1, wherein:

wherein the influence rule is:

Wherein the relationship degree weight is the importance degree value of the influence relationship among the subjects,

the calculation of the influence force between the main bodies comprises the following steps:

step A1, calculating the data to be analyzed by using a correlation diagram algorithm based on the knowledge diagram to obtain the influence of a single main body of all main bodies in the knowledge diagram and the association degree among the main bodies;

step a2, sequentially calculating the degree of influence between the main body and each of the other main bodies in the supply chain including the main body according to the formula (1) to obtain the inter-main-body risk value between each of the other main bodies corresponding to all the main bodies in the supply chain.

7. The supply chain risk management system of claim 1,

wherein the risk value calculation rule comprises the steps of:

step C1, analyzing the data to be analyzed by using a data mining method to obtain a plurality of characteristic parameters of the main body;

and step C2, inputting all the characteristic parameters into a pre-trained machine learning model to obtain the initial risk value of each subject.

8. The supply chain risk management system of claim 1, further comprising:

a weight storage module for storing the weight of the object,

wherein the weight storage module stores a plurality of influence weight values corresponding to each of the security elements and corresponding to degrees of risk importance involved by the respective subjects in the security elements,

9. The supply chain risk management system of claim 1, wherein:

wherein the pretreatment specifically comprises the following steps:

step S1, counting the history tracing information, completing the history tracing information with less missing items, and deleting the history tracing information with more missing items;

step S2, cross fusion and comparison are carried out on the history traceability information through a machine learning algorithm to obtain cross information;

step S3, screening the cross information, and modifying or deleting the wrong cross information to obtain the data to be stored;

step S4, performing association processing on the environment information to obtain associated environment data;

step S5, storing the data to be stored and the data related to the relational analysis in the associated environment data to the atlas storage module, storing the data not related to the relational analysis as the data to be analyzed to the traditional relational database,

10. The supply chain risk management system of claim 1, wherein:

the supply chain information storage module is a blockchain associated with all main bodies in the supply chain, and the blockchain is used for automatically acquiring related filing information, authentication information, supply chain related basic information of the main bodies and supply chain related public opinion information and environmental public opinion information on each website, and correspondingly storing the related filing information, the authentication information and the supply chain related basic information as the history traceability information and the supply chain related public opinion information and the environmental public opinion information as the environmental information.