A supply chain risk assessment and prevention system

Abstract

The application discloses a kind of supply chain risk assessment and prevention and control system, including resource acquisition and identification module, risk assessment module;Resource acquisition and identification module are used to collect external data, and data are extracted, converted and cleaned, form labeled data to be stored and managed;Risk assessment module is used to build risk model, then according to the calculation result of basic rule and calling risk model, with risk assessment result guiding production decision, risk prevention and control are carried out.The application uses CVaR conditional risk value, and carries out demand identification based on the customization of user order, resolves multi-source heterogeneous data according to business requirements, and the supply chain with overall risk value being too high can realize node deletion and replacement, with high reusability and high flexibility, multi-dimensional risk modeling is carried out according to the characteristics of three-level supply chain, and the supply chain risk value is obtained, to guide and make decisions for enterprise supply chain risk and supply chain management.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of computer big data, and particularly relates to a supply chain risk assessment and prevention system. BACKGROUND

[0002] The biggest challenge for enterprise development in today's e-commerce era is that large-scale manufacturing must be changed into a large-scale customization mode, that is, from originally manufacturing products first and then finding users to selecting users first and then manufacturing products. The traditional enterprise "production-stock-sales" mode must be correspondingly changed into a user-driven "on-demand" mode. Supply chain management plays an increasingly important role under large-scale customization of enterprises. The participants involved in the supply chain are numerous and interlocked with each other, and any sudden problem in any link can bring adverse effects to the upstream and downstream of the entire supply chain. Therefore, the supply chain system has great uncertainty and vulnerability, and is vulnerable to various risks.

[0003] At present, most of the existing researches are supply chain operation and risk assessment management between single enterprises, and there are relatively few supply chain risk assessment systems under multi-dimensional enterprise supply chain risk modeling and multi-level supply chain and multi-enterprise mode. SUMMARY

[0004] The application aims to provide a supply chain risk assessment and prevention system which uses CVaR conditional risk value to identify customization demand of user orders, analyzes multi-source heterogeneous data according to business demand, performs multi-dimensional risk modeling according to the characteristics of a three-level supply chain, and realizes node deletion and replacement for a supply chain with an overall risk value that is too high, and has high reusability, high flexibility and easy expansion.

[0005] To achieve the above object, the technical scheme adopted by the application is as follows:

[0006] A supply chain risk assessment and prevention system comprises a resource acquisition and identification module and a risk assessment module, characterized in that the resource acquisition and identification module is used to acquire external data, and perform extraction, conversion and cleaning on the data to form labeled data for storage and management; the risk assessment module is connected with the resource acquisition and identification module, and is used to construct a risk model according to the data processed by the resource acquisition and identification module, then guide production decision-making with a risk assessment result according to a basic rule and a calculation result of the risk model, and perform risk prevention and control;

[0007] The resource acquisition and identification module comprises a data query unit, a resource label management unit and a data labeling storage unit.

[0008] The data query unit is connected with the data tagging storage unit, and is used for querying the tagged data of the enterprise from the HDFS by taking the enterprise registration ID as a primary key after data storage.

[0009] The resource tag management unit is connected with the data tagging storage unit, and is used for continuously extending and expanding the tagged data, and expanding and modifying the rules of data extraction, conversion and cleaning.

[0010] The risk assessment module comprises a risk model construction unit and a risk decision unit.

[0011] The risk decision unit is connected with the risk model construction unit, and is used for configuration and management of basic rules and risk models, and according to triggering of real-time events of the system, through calling of the rule library unit, a plurality of supply chain nodes are spliced and combined to form a complete supply chain system, so that matching of risk prevention and control rules and execution of decision actions are completed.

[0012] The risk model construction unit is connected with the resource collection and recognition module, and is used for risk model modeling and configuration through online calculation and offline calculation of real-time supply chain data according to the data processed by the resource collection and recognition module.

[0013] The risk decision unit comprises a risk data injection unit, a rule library unit, a rule management unit and a supply chain model combination unit.

[0014] The rule management unit is connected with the rule library unit, and is used for basic rule configuration and risk model parameter configuration.

[0015] The rule library unit comprises basic rules and risk models.

[0016] The basic rules are hard rules without logic processing, and are a kind of rules directly defined by users or enterprises.

[0017] The risk model is injected by supply chain multidimensional data, mainly based on business logic, adjusted according to enterprise capability, product capability, production capability and distribution capability, and through calling of risk model algorithm, output actions of the supply chain are decided, the feasibility of an order of an enterprise user is evaluated, and rapid judgment and landing execution of the order are supported.

[0018] The supply chain model combination unit is connected with the rule library unit, and is used for calling enterprise node models in the rule library unit to splice and combine upstream and downstream supply chains in a way of dragging nodes, deleting or replacing supply chain nodes can be clicked, flexible combination is carried out, a complete supply chain network structure can be configured, and a quantitative risk value is decided by the risk model to guide production decision.

[0019] The tag data preferably includes business data, production data, credit data and other external data.

[0020] The business data is preferably associated with the same enterprise from different channels by a unique primary key (enterprise registration ID) to initiate upstream and downstream transactions, form a complete historical transaction sequence, realize supply chain resource integration, and include enterprise internal resources, supporting enterprise resources and other resources in the region.

[0021] The production data includes enterprise raw materials, merchants, equipment, orders, inventory, production, distribution, cost, quality, benefit and other multi-dimensional product capability resources, supply chain distribution capability resources and manufacturing capability resources.

[0022] The credit data includes enterprise credit, payment method and credit limit.

[0023] The other external data is connected through external interfaces with other enterprise platforms or HTTP external data.

[0024] The risk model construction unit preferably comprises:

[0025] The risk model is defined as enterprise capability, product capability, production capability and distribution capability, which are used as templates.

[0026] The risk model calculates the supply chain risk by CVaR. The CVaR value of a supply chain node enterprise means that the loss of the enterprise at a certain supply chain node within a certain time period exceeds a given VaR value at a certain probability level.

[0027] The formula is:

[0028] CVaR β =VaR β +E{f(x,y)-VaR β |f(x,y)≥VaR β} Formula 1

[0029] In formula 1,

[0030] x=(x1,x2,x3,…x n ) M , represents the weight of the multi-dimensional influencing factors of enterprise capability, product capability, production capability and distribution capability in the enterprise supply chain definition model;

[0031] y=(y1,y2,y3,…y n ) M , represents the weight of the product value affected by different risk factors in the enterprise or supply chain;

[0032] f(x, y) represents the expected value loss function of the product value in the enterprise supply chain within a certain holding period;

[0033] VaR β represents the supply chain node risk value of the enterprise at the probability level β, that is, the upper limit of the possible loss;

[0034] Online calculation, through the home appliance enterprise ID to query the historical transaction sequence, find out the enterprise historical transaction on the supply chain of multiple raw material suppliers, multiple levels of distributors, and receive real-time enterprise data; Through the message queue component to inject the stream computing engine, simulate the single node risk of raw material suppliers, manufacturers, and multiple levels of distributors in the supply chain;

[0035] Offline calculation, mainly to learn the historical experience data of the supply chain, complete the characteristic value calculation through scheduling the big data distributed cluster, including calculating the influence weight of the point node enterprise ability, product ability, production ability, and distribution ability in the supply chain risk, and the weight of product value;

[0036] After the data calculation, the XML resource description template file of class, attribute, and value definition is generated and stored in the distributed cache; The risk model will be tested in the gray environment, simulating the running situation of the single node production environment of the supply chain enterprise;

[0037] The risk feature management is responsible for the optimization adjustment of the parameters of the risk model, the whole life cycle management and the continuous updating and iteration of the model; After continuous iteration and optimization, the module transmits data with the risk decision unit through the self-defined Java interface component, and stores the risk model of the single node of the enterprise in the supply chain into the rule library unit.

[0038] Preferably, the risk decision unit specifically comprises:

[0039] The risk decision unit is injected with multi-dimensional data of the supply chain, mainly based on business logic, and is adjusted according to enterprise ability, product ability, production ability, and distribution ability. The output action of the supply chain is decided by calling the risk model algorithm, the feasibility of the enterprise user order is evaluated, and the rapid judgment and landing execution of the order are supported;

[0040] The risk model connects raw material suppliers, manufacturers, transportation and distribution, downstream distributors, and forms a joint network system, involving a three-level supply chain model of multiple suppliers, manufacturers, and multiple distributors; Due to the subadditivity of CVaR, the risk value Risk of the whole supply chain of a certain manufacturer M is as follows:

[0041]

[0042] Regarding formula 2, wherein,

[0043] Ut represents the tthupstream supplier;

[0044] M’ represents the mthcore manufacturer;

[0045] Dr represents the rthdownstream distributor;

[0046] λ m CVaR β,M' represents the risk value of the manufacturer itself;

[0047] represents the risk value of the core manufacturer and the upstream supplier trading to itself;

[0048] represents the risk value of the core manufacturer and the downstream distributor trading to itself;

[0049] CVaR β,M' respectively represents the CVaR value of the supplier, the manufacturer and the distributor under a given probability level β;

[0050] and respectively represents the risk influence factor of the upstream supplier and the downstream distributor on the core manufacturer.

[0051] Preferably, the risk decision unit, in particular:

[0052] When all the conditions of a basic rule are met, the subsequent risk prevention and control actions need to be executed; according to the risk model prediction, combined with the risk influence factors of different upstream and downstream enterprises, the overall risk value of the supply chain is calculated; different levels of risk model prediction results will execute corresponding actions, mainly including three risk levels, the first level (risk value less than 60%) includes abnormal situation warning prompt of possible or already appeared in each link of the supply chain, and the relevant personnel are notified and guided to handle, this level does not affect the normal operation of the supply chain; the second level (risk value is 60%-80%) will affect the operation stability of the supply chain, and is easy to produce adverse consequences, and the management personnel may need to replace individual nodes, and need to be handled as soon as possible; the third level (risk value greater than 80%) will directly make the supply chain stop operation, and the management personnel need to delete or replace multiple nodes, or directly stop the operation of the supply chain, which is the most strict intervention measure.

[0053] Preferably, the enterprise capability includes credit, payment, cooperation and black and white list;

[0054] The product capability includes confidence, CPK, quality other indicators, cost, profit and time period;

[0055] The production capacity includes production order decomposition, work order, production scheduling, raw material inventory, processing material inventory, other material inventory, equipment operation, equipment life cycle, equipment maintenance and production technology assembly;

[0056] The distribution capacity includes upstream and downstream logistics and information flow.

[0057] Preferably, the risk data injection unit adopts a parameter-unified XML structure.

[0058] The present application has the following beneficial effects:

[0059] By simulating user customization requirements, the supply chain risk model is defined as enterprise capacity, product capacity, production capacity and distribution capacity, four categories of capacity including respective sub-branches, so that the model definition is clearer, and template reuse is supported, and modeling is more convenient.

[0060] The risk assessment module models the risk of a single node in the upstream and downstream of the supply chain, then combines the nodes of the upstream and downstream for the manufacturer, and for the supply chain with too high overall risk value, the nodes can be deleted and replaced, so that the flexibility is higher, and the cost of the supply chain is reduced.

[0061] In the risk modeling, the CVaR conditional risk value is introduced, demand recognition is carried out based on user order customization, multi-source heterogeneous data is analyzed according to business requirements, multi-dimensional risk modeling is carried out according to the characteristics of the three-level supply chain, and the supply chain risk value is obtained, so as to guide and make decisions for enterprise supply chain risk and supply chain management. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 A supply chain risk assessment and prevention and control system structure diagram of the present application;

[0063] Figure 2 A risk model schematic diagram of the present application; DETAILED DESCRIPTION

[0064] The present application will be further described below in combination with the drawings and embodiments:

[0065] Embodiment 1

[0066] As Figure 1 , 2As shown, the present application provides a supply chain risk assessment and prevention system, comprising a resource collection and identification module 1, a risk assessment module 2; the resource collection and identification module 1 is used to collect external data, and the data is extracted, converted and cleaned to form labeled data for storage and management; the risk assessment module 2 is connected with the resource collection and identification module 1, and is used to construct a risk model 13 according to the data processed by the resource collection and identification module 1, and then according to the calculation results of the risk model 13 and the basic rules 12, the production decision is guided by the risk assessment results, and the risk prevention is carried out.

[0067] The resource collection and identification module 1 belongs to the bottom layer of the system, comprising a data query unit 4, a resource label management unit 5 and a data labeling storage unit 3.

[0068] In the production of customized household appliances, the resource collection and identification module 1 module realizes the storage of massive production data, business data, credit data and other external data of household appliance production and upstream and downstream enterprises by building a Hadoop ecology.

[0069] The resource collection and identification module 1 first queries whether there is enterprise historical data from the HDFS file storage system through the unique primary key (enterprise registration ID), if there is, the necessary historical data update is carried out, if there is not, the production, business, credit and other data are obtained through the open data interface of the enterprise.

[0070] After obtaining the data, the data is extracted, converted and cleaned by the data ETL component to form labeled data for data management, and the labeled data formed includes business data, production data, credit data and other external data.

[0071] The business data is associated and matched together by the unique primary key (enterprise registration ID) to form a complete historical transaction sequence, realizing the integration of supply chain resources, including enterprise internal resources, supporting enterprise resources and other resources in the region. The production data is from the multi-dimensional product capability resources, supply chain distribution capability resources and manufacturing capability resources of raw materials, merchants, equipment, orders, inventory, production, distribution, cost, quality and benefit. The credit data includes comprehensive monitoring of enterprise credit investigation, payment method and credit limit. Other risk data is connected through external interfaces with other enterprise platforms or HTTP external data.

[0072] The data query unit 4 is connected with the data labeling storage unit 3: after data storage, the business data, production data, credit data and other external data of the enterprise can be queried from the HDFS by taking the enterprise registration ID as the primary key.

[0073] The resource tagging management unit 5 is connected to the data tagging storage unit 3. The tagged data is dynamic, allowing the system to continuously extend and expand the tagged data as the basic risk data is updated, and to expand and modify the rules for data extraction, transformation and cleaning.

[0074] Resource acquisition and identification module 1 runs a Java-based client program, which communicates with risk assessment module 2 and uploads data files through the REST-based HDFS API.

[0075] Risk assessment module 2 includes risk model building unit 6 and risk decision-making unit 7.

[0076] The risk model building unit 6 is connected to the resource acquisition and identification module 1 and is used to model and configure the risk model based on the data processed by the resource acquisition and identification module 1 through online and offline calculations of real-time supply chain data.

[0077] Risk assessment module 2, by simulating user-customized needs, defines risk model 13 as enterprise capabilities, product capabilities, production capabilities, and distribution capabilities, using this as a template. These four categories of capabilities each contain their own sub-branches, as detailed below. Figure 2 As shown. Enterprise capabilities include creditworthiness, payment, cooperation, and blacklists / whitelists; product capabilities include confidence level, CPK, other quality indicators, cost, profit, and time cycle; production capabilities include production order breakdown, work orders, production scheduling, raw material inventory, processed material inventory, other material inventory, equipment operation status, equipment life cycle, equipment maintenance, and production technology assembly; distribution capabilities include upstream and downstream logistics and information flow.

[0078] Risk Model 13 for a Single Node in the Home Appliance Supply Chain:

[0079] Risk Model 13 uses CVaR to calculate supply chain risk. The CVaR value of a supply chain node enterprise means, under a certain probability level, the expected amount of loss of a certain supply chain node of an enterprise within a certain period of time exceeds a given VaR value.

[0080] The formula is:

[0081] CVaR β =VaR β +E{f(x,y)-VaR β |f(x,y)≥VaR β Equation 1

[0082] Regarding equation 1,

[0083] x = (x1, x2, x3, ... x n ) M, which represents the weight of the multi-dimensional influencing factors of enterprise capability, product capability, production capability, and distribution capability in the enterprise supply chain definition model;

[0084] y = (y1, y2, y3, … y n ) M , which represents the weight of the product value influenced by different risk factors in the enterprise or supply chain;

[0085] f(x, y) represents the expected value loss function of the product value in the enterprise supply chain within a certain holding period;

[0086] VaR β represents the supply chain node risk value of the enterprise at a probability level β, i.e., the upper limit of possible loss.

[0087] Online calculation finds out multiple raw material suppliers upstream and multiple levels of distributors downstream of the enterprise in the supply chain by querying the historical transaction sequence of the home appliance enterprise ID and receiving real-time enterprise data. Through the message queue component, the stream computing engine is injected to simulate the single-node risk of raw material suppliers, manufacturers, and multiple levels of distributors in the supply chain.

[0088] Offline calculation mainly learns from historical experience data of the supply chain and completes feature value calculation through scheduling of a big data distributed cluster, including calculation of the influence weight of point node enterprise capability, product capability, production capability, and distribution capability in the supply chain risk, and the weight of product value.

[0089] After data calculation, an XML resource description template file is generated for class, attribute, and value definition and stored in a distributed cache. The risk model 13 is tested in a gray environment to simulate the operation of a single-node production environment of a supply chain enterprise. The risk feature management is responsible for optimization adjustment of parameters of the risk model 13, full life cycle management, and continuous updating and iteration of the model. After continuous iteration and optimization, the risk model construction unit 6 transmits data to the risk decision unit 7 through a self-defined Java interface component, and finally the risk model 13 of a single node of an enterprise in the supply chain is stored in the rule library unit 9.

[0090] The risk decision unit 7 includes a risk data injection unit 8, a rule library unit 9, a rule management unit 10, and a supply chain model combination unit 11;

[0091] The risk decision unit 7 is connected with the risk model construction unit 6, and is used for configuration and management of the basic rule 12 and the risk model 13. According to triggering of a system real-time event, a plurality of supply chain nodes are spliced and combined through calling of the rule library unit 9, so that a complete supply chain system is formed, and matching of the risk prevention and control rule and execution of the decision action are completed. In order to facilitate injection of risk data of a heterogeneous system, an entry parameter of the risk data injection unit 8 is uniformly defined in an XML structure. The rule library unit 9 is composed of the specific basic rule 12 and the risk model 13, and jointly makes decisions. The specific rule is as follows:

[0092] The basic rule 12 is a hard rule without logic processing, and is a kind of rule whose rule content is directly defined by a user or an enterprise. If a certain node on an upstream or downstream supply chain does not meet a condition, the certain node can be directly deleted or replaced. Taking a home appliance enterprise as an example, the basic rule 12 can be an upper limit of daily output, an upper limit of production cost, an enterprise credit problem and the like.

[0093] An enterprise can independently add or modify a rule, and can publish a new version. The basic rule 12 has the advantage of strong timeliness of the rule and low requirement.

[0094] The risk model 13 is injected with multi-dimensional data of a supply chain, and is mainly based on business logic. The risk model 13 is adjusted according to enterprise capability, product capability, production capability and distribution capability, and decides an output action of the supply chain through calling of a risk model algorithm, evaluates feasibility of an enterprise user order, and supports rapid judgment and landing execution of the order.

[0095] The risk model 13 connects raw material suppliers, production manufacturers, transportation distributors, downstream distributors and the like to form a joint network system, and involves a three-level supply chain model of a plurality of suppliers, manufacturers and distributors. Due to subadditivity of CVaR, it is obtained that a risk value Risk of a whole supply chain of a certain manufacturer M is as follows:

[0096]

[0097] Regarding formula 2, wherein,

[0098] Ut represents a tth upstream supplier;

[0099] M' represents a mth core manufacturing enterprise;

[0100] Dr represents a rth downstream distributor;

[0101] λ m CVaR β,M' represents a risk value of the manufacturing enterprise itself;

[0102] represents a risk value brought to the core manufacturer by trade between the core manufacturer and the upstream supplier.

[0103] represents the risk value brought to the core manufacturer by the trade with the downstream distributor;

[0104] CVaR β,M' respectively represent the CVaR values of the suppliers, manufacturers and distributors at a given probability level β;

[0105] and respectively represent the risk influence factors of the upstream suppliers and downstream distributors on the core manufacturer.

[0106] The supply chain model combination unit 11 is connected with the rule library unit 9 and can be interacted through an html page, and is used to call the enterprise node model in the rule library unit 9 to splice and combine the upstream and downstream supply chains in a way of dragging nodes, and can click to delete or replace the supply chain nodes for flexible combination. Taking a home appliance enterprise as an example: the manufacturer can select the upstream supplier enterprise, and each supplier node needs to produce and deliver how much raw materials, the delivery date and the like; select the downstream distributor enterprise, and each distributor node delivers how much finished products and the price and the like. A complete supply chain network structure that can be configured is formed, and the risk value is quantified through model decision to guide production decision.

[0107] The rule management unit 10 is connected with the rule library unit 9, and is used for business personnel to configure and modify the basic rules 12 and the risk model 13 parameters through an html interactive page. In addition, the management of the enterprise black, white and gray list can also be added.

[0108] When all the conditions of a basic rule are met, the subsequent risk prevention and control actions need to be performed; according to the risk model prediction, the overall supply chain risk value is calculated in combination with the risk influence factors of different upstream and downstream enterprises; different levels of risk model prediction results will perform corresponding actions, mainly including three risk levels, the first level (risk value less than 60%) includes abnormal situation warning prompt that may or has appeared in each link of the supply chain, and notifies and guides the relevant personnel to handle, this level does not affect the normal operation of the supply chain; the second level (risk value is 60%-80%) will affect the operation stability of the supply chain, and is easy to produce adverse consequences, and the management personnel may need to replace individual nodes, and need to handle as soon as possible; the third level (risk value greater than 80%) will directly make the supply chain stop operation, and the management personnel need to delete or replace multiple nodes, or directly stop the operation of this supply chain, which is the most strict intervention measure.

[0109] The above merely describes preferred embodiments of the present application, but is not intended to limit the present application to other forms, and any person skilled in the art can make changes or modifications to the above disclosed technical contents into equivalent embodiments with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiments without departing from the technical solution content of the present application and according to the technical essence of the present application still belongs to the protection scope of the technical solution of the present application.

Claims

1. A supply chain risk assessment and prevention system, comprising a resource collection and identification module, a risk assessment module; characterized in that: The resource collection and identification module is configured to collect external data, and to extract, convert and clean the data to form labeled data for storage and management; the risk assessment module is connected to the resource collection and identification module, and is configured to construct a risk model according to the data processed by the resource collection and identification module, and then to guide production decisions according to a risk assessment result based on a basic rule and a calculation result of the risk model, and to perform risk prevention and control; The resource collection and identification module comprises a data query unit, a resource label management unit and a data labeling storage unit; The data query unit is connected to the data labeling storage unit, and is configured to query the labeled data of an enterprise from HDFS by taking an enterprise registration ID as a primary key after data storage; The resource label management unit is connected to the data labeling storage unit, and is configured to continuously extend and expand the labeled data, and to expand and modify the rules for data extraction, conversion and cleaning; The risk assessment module comprises a risk model construction unit and a risk decision unit; The risk decision unit is connected to the risk model construction unit, and is configured to configure and manage the basic rules and the risk model, to trigger a real-time event of the system, to perform splicing and combination on multiple supply chain nodes through calling of a rule library unit, to form a complete supply chain system, and to complete matching of risk prevention and control rules and execution of decision actions; The risk model construction unit is connected to the resource collection and identification module, and is configured to construct and configure a risk model according to the data processed by the resource collection and identification module through online calculation and offline calculation of real-time supply chain data; The risk decision unit comprises a risk data injection unit, a rule library unit, a rule management unit and a supply chain model combination unit; The rule management unit is connected to the rule library unit, and is configured to configure basic rules and risk model parameters; The rule library unit comprises basic rules and a risk model; The basic rules are hard rules that do not require logical processing, and are a type of rule whose content is directly defined by a user or an enterprise; The risk model is based on business logic, and is adjusted according to enterprise capability, product capability, production capability and distribution capability, and decides an output action of a supply chain by calling a risk model algorithm, evaluates the feasibility of an order of an enterprise user, and supports rapid judgment and landing execution of the order; The supply chain model combination unit is connected to the rule library unit, and is configured to call an enterprise node model in the rule library unit, to splice and combine an upstream and downstream supply chain in a manner of dragging a node, to click to delete or replace a supply chain node, to flexibly combine, to form a configurable complete supply chain network structure, and to guide production decisions by quantifying a risk value through the risk model; The risk model construction unit specifically comprises: The risk model is defined as enterprise capability, product capability, production capability and distribution capability, and is used as a template; The risk model calculates supply chain risks through CVaR. The CVaR value of the supply chain node enterprise means that the loss of the enterprise in a certain supply chain node exceeds a given VaR value in a certain period of time under a certain probability level; The formula is: CVaR β = VaR β + E{f(x,y) - VaR β |f(x,y) > VaR β} Equation 1 In formula 1, x = (x1, x2, x3, Λ x n ) M , represents the weight of the multi-dimensional influencing factors of enterprise capability, product capability, production capability, and distribution capability in the enterprise supply chain definition model. y = (y1, y2, y3, Λ y n ) M , represents the weight of the product value within the enterprise or supply chain due to the influence of different risk factors; f(x, y) represents the expected value loss function of the product value in the enterprise supply chain within a certain holding period; VaR β represents the value at risk of the supply chain node of the enterprise at the probability level β, i.e. the upper limit of the possible loss; Online calculation, through the home appliance enterprise ID query historical transaction sequence, find out the enterprise historical transaction on the supply chain of multiple raw material suppliers, multi-level distributors, and receive real-time enterprise data; through the message queue component injection flow calculation engine, simulate the single node risk of raw material suppliers, manufacturers, multi-level distributors in the supply chain; Offline calculation, learn from the historical experience data of the supply chain, and complete the characteristic value calculation through scheduling the big data distributed cluster, including calculating the influence weight of single node enterprise capability, product capability, production capability, and distribution capability in the supply chain risk, and the weight of product value; After data calculation, an XML resource description template file is generated, which defines classes, attributes, and values, and is stored in a distributed cache; the risk model is tested in a gray environment to simulate the operation of the single node production environment of the supply chain enterprise; The risk feature management is responsible for the optimization and adjustment of the parameters of the risk model, the whole life cycle management and the continuous updating and iteration of the model; after continuous iteration and optimization, the module transmits data to the risk decision unit through a self-defined Java interface component, and stores the risk model of the single node of the enterprise in the supply chain into the rule library unit. 2.The supply chain risk assessment and prevention system according to claim 1, characterized in that: The tagged data includes business data, production data, credit data and other external data. 3.The supply chain risk assessment and prevention system according to claim 2, characterized in that: The business data is associated and matched through a unique primary key to form a complete historical transaction sequence, realize the integration of supply chain resources, including enterprise internal resources, supporting enterprise resources, and other resources in the region. The production data includes multi-dimensional product capability resources, supply chain distribution capability resources, and manufacturing capability resources of raw materials, merchants, equipment, orders, inventory, production, distribution, cost, quality, and benefit of the enterprise; The credit data includes enterprise credit, payment method, and credit limit; The other external data is connected through external interfaces with other enterprise platforms or HTTP external data. 4.The supply chain risk assessment and prevention system according to claim 1, characterized in that: The risk decision unit specifically comprises: Based on business logic, the risk model is adjusted according to enterprise capability, product capability, production capability, and distribution capability, and the output action of the supply chain is decided by calling the risk model algorithm, so as to evaluate the feasibility of the enterprise user order and support the rapid judgment and landing execution of the order; The risk model connects raw material suppliers, manufacturers, transportation and distribution, downstream distributors and other joint network systems, involving a three-level supply chain model of multiple suppliers, manufacturers and distributors; due to the subadditivity of CVaR, the risk value Risk of the whole supply chain of a certain manufacturer M is as follows: In formula 2, Ut represents the tthupstream supplier; M' represents the mthcore manufacturer; Dr represents the rth downstream distributor; λ m CVaR β,M' represents the risk value of the manufacturing enterprise itself; represents the risk value that the core manufacturer trades with the upstream supplier to itself; represents the risk value that the core manufacturer and downstream distributor trade to themselves; respectively denote the CVaR values of the suppliers, manufacturers, distributors at a given probability level β; and respectively represent the risk impact factors of the upstream supplier and the downstream distributor on the core manufacturer.

5. The supply chain risk assessment and prevention system of claim 4, wherein: The risk decision unit, in particular: When all conditions of a basic rule are met, the subsequent risk prevention action needs to be executed; according to the risk model prediction, combined with the risk influence factors of different upstream and downstream enterprises, the overall risk value of the supply chain is calculated; different levels of risk model prediction results will execute corresponding actions, including three risk levels, the first level risk value is less than 60%, including warning and prompt for possible or existing abnormal situation of each link of the supply chain, notifying and guiding relevant personnel to handle, this level does not affect the normal operation of the supply chain; the second level risk value is 60%-80%, which will affect the operation stability of the supply chain, and is easy to produce adverse consequences, and the management personnel may need to replace individual nodes, and need to handle as soon as possible; the third level risk value is greater than 80%, which will directly make the supply chain stop operation, and the management personnel need to delete or replace multiple nodes, or directly stop the operation of the supply chain, which is the most strict intervention measure.

6. The supply chain risk assessment and prevention system of claim 5, wherein: The enterprise capability includes credit, payment, cooperation and black and white list; The product capability includes confidence, CPK, quality other indicators, cost, profit and time period; The production capability includes production order decomposition, work order, production scheduling, raw material inventory, processing material inventory, other material inventory, equipment operation, equipment life cycle, equipment maintenance and production technology assembly; The distribution capability includes upstream and downstream logistics and information flow.

7. The supply chain risk assessment and prevention system of claim 6, wherein: The risk data injection unit adopts a parameter-unified XML structure.

Images