Electronic system complete machine production plan risk assessment method, device, equipment and product

Abstract

The invention discloses an electronic system complete machine production plan risk assessment method, device, equipment and product, relates to the field of digital production management, and is used for improving the accuracy and interpretability of production plan risk assessment. According to the determined to-be-predicted risk type, the risk variable data of each risk type is collected and the data quality is evaluated; respectively carrying out dominant risk judgment and recessive risk prediction on the risk variable data; evaluating the initial risk of each risk type based on the dominant risk and the recessive risk; and based on the data quality, the dominant risk and the implicit risk, respectively calculating a dominant risk weight, an implicit risk weight and a preliminary risk weight of each risk type, and respectively calculating a comprehensive risk of each risk type. According to the method, the comprehensiveness, objectivity and accuracy of the evaluation result are remarkably improved, the risk calculation logic is traceable, and the interpretability of the evaluation result is high.

Description

Technical Field

[0001] The present invention relates to the technical field of digital production management, and in particular to a method, device, equipment and product for risk assessment of the overall production plan of an electronic system. Background Art

[0002] A typical production mode of the overall electronic system with multiple varieties, variable batches and discrete types has the characteristics of high quality requirements, large delivery volume, tight cycle and complex process. In recent years, the internal and external environments faced by the production and manufacturing of the overall electronic system have increasingly shown the characteristics of volatility, uncertainty, complexity and ambiguity. Therefore, the risks faced in its production process are becoming more complex in terms of quantity and type, and these risks will seriously affect the achievement of the delivery period target. Therefore, production plan risk management (such as risk assessment, risk response, etc.) has increasingly become an important and indispensable part of production.

[0003] For the production scenario of the overall electronic system, the current production plan risk assessment methods mainly have the following deficiencies: (1) Over-reliance on qualitative methods. Many enterprises mainly rely on the experience judgment of business personnel for risk assessment, lacking data support, resulting in strong subjectivity of the assessment results, and insufficient accuracy and consistency.

[0004] (2) Insufficient application and poor effect of quantitative methods. Quantitative methods have high requirements for data quality, and it is difficult to provide reliable decisions when data is insufficient or inaccurate, affecting the reliability and effectiveness of risk assessment. In addition, the evaluation results of the model generally lack interpretability.

[0005] (3) Single method and lack of comprehensive evaluation. Often a single method is adopted, and multiple evaluation techniques are not combined, and various risks cannot be comprehensively covered. Summary of the Invention

[0006] The object of the present invention is to provide a method, device, equipment and product for risk assessment of the overall production plan of an electronic system to improve the accuracy and interpretability of the risk assessment of the overall production plan of an electronic system for all or part of the above problems.

[0007] The technical solution adopted by the present invention is as follows: A method for risk assessment of the overall production plan of an electronic system, which includes: According to the determined risk types to be predicted, collect the risk variable data of each risk type respectively and evaluate the data quality; Based on a preset risk inference rule set, conduct explicit risk determination on each risk variable data; Based on a pre-trained risk prediction model, conduct implicit risk prediction on each risk variable data; Based on the explicit risk and implicit risk, evaluate the preliminary risk of each risk type; Calculate the explicit risk weight, implicit risk weight, and preliminary risk weight for each risk type based on data quality, explicit risk, and implicit risk respectively; Calculate the comprehensive risk for each risk type based on the explicit risk weight, implicit risk weight, and preliminary risk weight respectively.

[0008] Furthermore, the explicit risk determination for each risk variable data based on a preset risk inference rule set includes: Calculate the rule coverage rate for each risk variable data based on a preset risk inference rule set respectively.

[0009] Furthermore, the implicit risk prediction for each risk variable data based on a pre-trained risk prediction model includes: Perform predictions on each risk variable data based on a pre-trained risk prediction model respectively to obtain the risk occurrence probability and confidence level for each risk type.

[0010] Furthermore, the preliminary risk assessment based on explicit risk and implicit risk includes: Based on the constructed knowledge base, match the explicit risk and implicit risk of each risk type to obtain a preliminary risk conclusion; the knowledge base contains judgment conditions and judgment conclusions for judging explicit risk and implicit risk.

[0011] Furthermore, the calculation of the explicit risk weight, implicit risk weight, and preliminary risk weight for each risk type based on data quality, explicit risk, and implicit risk respectively includes: Calculate the explicit risk weight, implicit risk weight, and preliminary risk weight for each risk type respectively according to the following method: Explicit risk weight = 0.3 × rule coverage rate + 0.2 × data quality; Implicit risk weight = 0.4 × confidence level + 0.1 × data quality; Preliminary risk weight = 1 - explicit risk weight - implicit risk weight.

[0012] Furthermore, the calculation of the comprehensive risk for each risk type based on the explicit risk weight, implicit risk weight, and preliminary risk weight respectively includes: Calculate the comprehensive risk for each risk type respectively according to the following method: Comprehensive risk = (explicit risk contribution × explicit risk weight) + (risk occurrence probability × implicit risk weight) + (preliminary risk contribution × preliminary risk weight); Among them, the explicit risk contribution is obtained by quantifying the explicit risk, and the preliminary risk contribution is obtained by quantifying the preliminary risk conclusion.

[0013] Furthermore, the method further includes: Output the comprehensive risk of each risk type and the calculation process of each comprehensive risk.

[0014] On the other hand, the present invention also provides an electronic system overall production plan risk assessment device, which includes: A first unit for respectively collecting risk variable data of each risk type according to the determined risk type to be predicted and evaluating the data quality; A second unit for making explicit risk judgments on the risk variable data based on a preset risk inference rule set; A third unit for making implicit risk predictions on the risk variable data based on a pre-trained risk prediction model; A fourth unit for evaluating the preliminary risk of each risk type based on the explicit risk and the implicit risk; A fifth unit for respectively calculating the explicit risk weight, the implicit risk weight and the preliminary risk weight of each risk type based on the data quality, the explicit risk and the implicit risk; A sixth unit for respectively calculating the comprehensive risk of each risk type based on the explicit risk weight, the implicit risk weight and the preliminary risk weight.

[0015] On the other hand, the present invention also provides an electronic system overall production plan risk assessment device, including a processor and a storage medium, the storage medium stores computer instructions, the processor is connected to the storage medium and is used to run the computer instructions to execute the above-mentioned electronic system overall production plan risk assessment method.

[0016] On the other hand, the present invention also provides a computer program product, including a computer program, when the computer program is run by a processor, it can execute the above-mentioned electronic system overall production plan risk assessment method.

[0017] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are: The electronic system overall production plan risk assessment solution provided by the present application evaluates the risk variable data from the explicit dimension and the implicit dimension respectively for each risk type, comprehensively considers the explicit risk and the implicit risk, adaptively selects a more reliable information source by designing a dynamic risk weight, can effectively avoid the conflict of evaluation results between different evaluation technologies while being able to integrate the advantages of different evaluation technologies, and at the same time avoid misjudgment caused by poor data quality of the collected risk variable data, significantly improving the comprehensiveness, objectivity and accuracy of the evaluation results, the risk calculation logic is traceable, and the interpretability of the evaluation results is strong. Description of the Drawings

[0018] The present invention will be described by way of examples with reference to the drawings, where: Figure 1 It is a flowchart of the risk assessment method for the overall production plan of an electronic system provided by an embodiment of the present application.

[0019] Figure 2 It is a schematic diagram of the risk index system for the overall production plan of an electronic system in an embodiment of the present application.

[0020] Figure 3 It is an architecture diagram of the risk assessment device for the overall production plan of an electronic system provided by an embodiment of the present application.

[0021] Figure 4 It is a structural diagram of the risk assessment device for the overall production plan of an electronic system provided by an embodiment of the present application. Detailed implementation manners

[0022] All features disclosed in this specification, or all steps in the disclosed methods or processes, except for mutually exclusive features and / or steps, can be combined in any manner.

[0023] Any feature disclosed in this specification (including any additional claims, abstract) can be replaced by other equivalent or similar-purpose alternative features unless specifically stated. That is, unless specifically stated, each feature is only an example of a series of equivalent or similar features.

[0024] Aiming at the problems of the current production plan risk assessment method being insufficiently comprehensive, subjective, overly dependent on high-quality data, and having poor interpretability, embodiments of the present application provide a risk assessment method, device, equipment, and product for the overall production plan of an electronic system, aiming to effectively improve the accuracy and interpretability of the risk assessment of the overall production plan of an electronic system.

[0025] As Figure 1 shown, the risk assessment method for the overall production plan of an electronic system provided by an embodiment of the present application includes the following processes: S1. According to the determined risk types to be predicted, collect the risk variable data of each risk type respectively and evaluate the data quality.

[0026] For the risk variable data to be collected for various risk types, according to the pre-constructed risk index system, the risk variable entries included in each risk type can be selected respectively as the attributes to be collected for collection.

[0027] For the links involved in the compilation of the overall production plan of an electronic system, in some feasible implementation manners, it is designed as Figure 2The risk index system for the six risk types shown includes: material risk, technical status risk, quality risk, cycle risk, resource risk, and support domain risk. Each risk type further includes one or more risk variable items respectively. For example, the risk variables involved in material risk are the status of historical stock materials, whether there is a replacement or model change, whether there is a shortage of equipment and supplies, etc.; the risk variables involved in quality risk are the historical average repair cycle, historical average failure rate, frequency of process suspension, etc.; the risk variables involved in resource risk are whether there is a need for key resources, whether capacity building lags behind, whether the manpower meets the requirements, whether the workstations meet the requirements, etc.

[0028] Based on the actual situation of the electronic system whole-machine production project, determine the risk types to be predicted. For example, in a certain electronic system whole-machine production project, the determined risk types include: 1) Material risk: The upstream department of the supply chain may not be able to supply materials on time; 2) Quality risk: The product failure rate is relatively high; 3) Resource risk: Resources may not be able to meet the demand.

[0029] According to the risk index system designed above, collect the risk variable data of material risk, quality risk, and resource risk respectively.

[0030] After the risk variable data of each risk type undergoes data preprocessing (such as data cleaning, outlier handling, data standardization, etc.), the data quality of the risk variable data of each risk type is determined by a pre-trained data quality assessment model or manually, and a data quality score for the risk variable data of each risk type is given. The basis for data quality assessment can be the integrity and accuracy of the risk variable data, that is, the data quality of the risk variable data is evaluated from these two dimensions. For example, for the three determined risk types above, the data quality assessment results shown in Table 1 are obtained.

[0031] Table 1 Data Quality Scoring Table

[0032] S2. Conduct explicit risk determination on each risk variable data based on a preset risk inference rule set.

[0033] The risk inference rule set covers all risk types. Taking the six risk types designed above as an example, the risk inference rule set includes risk inference rules for the six risk types. According to the determined risk types to be predicted, matching analysis is carried out in the risk inference rules of the corresponding risk types. Table 2 shows some examples of risk inference rules.

[0034] Table 2 Example Table of Risk Inference Rule Set

[0035] As an alternative implementation, the method for explicit risk determination includes: Based on a preset risk inference rule set, calculate the rule coverage rate of each risk variable data respectively. In addition, record whether each type of risk data triggers "high risk" (such as whether it triggers "high material risk", "high technical status risk", "high quality risk", etc.), that is, whether it triggers the risk inference rule corresponding to "high risk". For each type of risk, its rule coverage rate is represented by the ratio of the number of risk inference rules triggered by its risk variable data in the risk inference rule set to the total number of risk inference rules designed for this type of risk, expressed as: .

[0036] In the formula, R1 represents the rule coverage rate, r represents the number of triggered risk inference rules, and R represents the total number of risk inference rules.

[0037] According to the actually collected risk variable data, obtain the explicit risk assessment result shown in Table 3.

[0038] Table 3 Explicit Risk Table

[0039] S3. Conduct implicit risk prediction on each risk variable data based on a pre-trained risk prediction model.

[0040] The risk prediction model is used to predict the probability of risk occurrence. The risk prediction model adopts a machine learning algorithm and is trained using pre-collected historical risk variable data and corresponding risk labels (occurred or not occurred). For each type of risk, train a risk prediction model for this type of risk respectively for implicit risk prediction.

[0041] As an alternative implementation, the method for predicting implicit risks includes: Based on a pre-trained risk prediction model, conduct predictions on each risk variable data respectively to obtain the risk occurrence probability and confidence level of each type of risk.

[0042] According to the actually collected risk variable data of 3 types of risks, obtain the implicit risk prediction result shown in Table 4.

[0043] Table 4 Implicit Risk Prediction Result Table

[0044] S4. Evaluate the preliminary risk of each type of risk based on explicit risk and implicit risk.

[0045] In this step, the evaluation results of explicit risks and the prediction results of implicit risks are integrated to conduct a preliminary assessment of the comprehensive risk. As an alternative implementation method, the methods for evaluating the preliminary risk include: Based on the constructed knowledge base, the preliminary risk conclusion is obtained by matching the explicit risks and implicit risks of each risk type. The knowledge base contains the judgment conditions and judgment conclusions for judging explicit risks and implicit risks. For a certain risk type, if its explicit risk and implicit risk simultaneously meet a certain judgment condition, the judgment conclusion corresponding to this judgment condition is the preliminary risk conclusion.

[0046] For example, assume that one of the judgment conditions and judgment conclusions in the knowledge base is: IF (risk occurrence probability > 0.8 AND risk inference rule set not triggered) THEN initiate manual review.

[0047] As mentioned in some of the previous embodiments, the determination results of explicit risks include the rule coverage rate, and the prediction results of implicit risks include the risk occurrence probability and confidence level. According to the actually collected risk variable data, the preliminary risk conclusion shown in Table 5 is obtained.

[0048] Table 5 Preliminary Risk Conclusion Evaluation Table

[0049] S5. Calculate the explicit risk weight, implicit risk weight, and preliminary risk weight of each risk type based on data quality, explicit risk, and implicit risk respectively.

[0050] Each risk weight is a function of data quality, the rule coverage rate of explicit risk, and the confidence level of implicit risk, that is: Risk weight = f(data quality, rule coverage rate, confidence level), where f(data quality, rule coverage rate, confidence level) represents a risk weight calculation function with data quality, rule coverage rate, and confidence level as input data.

[0051] Specifically, in some feasible implementation methods, the calculation methods for each risk weight include: Calculate the explicit risk weight, implicit risk weight, and preliminary risk weight of each risk type respectively according to the following methods: Explicit risk weight = 0.3 × rule coverage rate + 0.2 × data quality; Implicit risk weight = 0.4 × confidence level + 0.1 × data quality; Preliminary risk weight = 1 - explicit risk weight - implicit risk weight.

[0052] In the above formula, the calculation coefficients of each risk weight are obtained by learning and fitting based on the risk variable data collected historically and can be adjusted appropriately. For example, each coefficient can be designed as a learnable parameter and obtained through training with the risk variable data collected historically.

[0053] The relevant parameters shown in Table 6 are calculated from the risk variable data actually collected.

[0054] Table 6 Input Data Table for Risk Weight Calculation Function

[0055] According to the calculation methods of the explicit risk weight, implicit risk weight, and preliminary risk weight provided in the above embodiments, the input data obtained from Table 6 is used to calculate the explicit risk weight, implicit risk weight, and preliminary risk weight shown in Table 7.

[0056] Table 7 Risk Weight Table

[0057] S6. Based on the explicit risk weight, implicit risk weight, and preliminary risk weight, calculate the comprehensive risk of each risk type respectively.

[0058] On the basis of dynamically calculating the explicit risk weight, implicit risk weight, and preliminary risk weight, based on the explicit risk weight, implicit risk weight, and preliminary risk weight, according to the explicit risk, implicit risk, and preliminary risk conclusions, the comprehensive risk of each risk type can be calculated respectively.

[0059] As an optional implementation method, the method for calculating the comprehensive risk includes: Comprehensive risk = (Explicit risk contribution × Explicit risk weight) + (Risk occurrence probability × Implicit risk weight) + (Preliminary risk contribution × Preliminary risk weight).

[0060] In the above formula, the explicit risk contribution is obtained by quantifying the explicit risk, and the preliminary risk contribution is obtained by quantifying the preliminary risk conclusion. Specifically, there are: In the explicit risk, if "high risk" is triggered, the explicit risk contribution is 1; if "high risk" is not triggered, the explicit risk contribution is 0. In the preliminary risk conclusion, if it is "high risk", the preliminary risk contribution is 1; if it is "initiate manual review", the preliminary risk contribution is 0.5; if it is "low risk" (or other than this), the preliminary risk contribution is 0.

[0061] Table 8 shows the data involved in calculating the comprehensive risk obtained from the risk variable data actually collected.

[0062] Table 8 Input Data Table for Comprehensive Risk Calculation

[0063] According to the comprehensive risk calculation method proposed above, the comprehensive risks of each risk type calculated from Table 8 are shown in Table 9.

[0064] Table 9 Comprehensive Risk Table

[0065] In addition, to further highlight the interpretability of the risk assessment results of the present application, in some alternative embodiments, the method further includes: Outputting the comprehensive risk of each risk type and the calculation process of each comprehensive risk. Among them, the output result can be converted into a natural language form for output.

[0066] In addition, in this step S6, the risk types can also be sorted by priority according to the calculated comprehensive risk to remind the risk types that need to be focused on. According to the comprehensive risks shown in Table 9, the finally output risk assessment results are shown in Table 10.

[0067] Table 10 Risk Assessment Result Table

[0068] In addition, in some alternative embodiments, the preset risk inference rule set and / or the pre-trained risk prediction model can also be updated and maintained to improve the accuracy of the evaluation results.

[0069] Specifically, use data samples (such as risk variable data and actual risk labels of one or more risk types) to test the above risk assessment method, obtain the comprehensive risk corresponding to the data sample, compare this comprehensive risk with the risk label (i.e., the actual risk value). If the error between the two exceeds the set threshold, it means that the evaluation deviation of the method of the present application is too large and needs to be updated. At this time, trigger the update of the risk inference rule set and / or trigger the retraining of the risk prediction model. The risk inference rule set can be updated by experts based on the latest statistically historical risk variable data and risk labels, and the risk prediction model can be retrained using the latest statistically historical risk variable data and risk labels.

[0070] According to the idea of the present application, an electronic system whole machine production plan risk assessment device is also provided in the embodiments of the present application, as Figure 3 shown, which includes: The first unit is used to collect the risk variable data of each risk type and evaluate the data quality according to the determined risk type to be predicted; The second unit is used to make an explicit risk determination for each risk variable data based on the preset risk inference rule set; A third unit for performing implicit risk prediction on each risk variable data based on a pre-trained risk prediction model; A fourth unit for evaluating the preliminary risks of each risk type based on explicit risks and implicit risks; A fifth unit for calculating the explicit risk weights, implicit risk weights, and preliminary risk weights of each risk type respectively based on data quality, explicit risks, and implicit risks; A sixth unit for calculating the comprehensive risks of each risk type respectively based on the explicit risk weights, implicit risk weights, and preliminary risk weights.

[0071] The specific data configured for each unit in the device can refer to the features designed in each step of the foregoing risk assessment method embodiments.

[0072] In addition, an embodiment of the present application also provides an electronic system overall production plan risk assessment device, as Figure 4 shown, including a processor and a storage medium. The storage medium stores computer instructions, and the processor is connected to the storage medium, for example, through a bus. The processor is used to run the computer instructions to execute the electronic system overall production plan risk assessment method of the foregoing embodiments.

[0073] In addition, an embodiment of the present application also provides a computer program product, including a computer program. When the computer program is run by a processor, it can execute the foregoing electronic system overall production plan risk assessment method.

[0074] The present invention is not limited to the foregoing specific embodiments. The present invention extends to any new feature or any new combination disclosed in this specification, as well as any new method or process step or any new combination disclosed.

Claims

1. A risk assessment method for the overall production plan of an electronic system, characterized in that, including: collecting risk variable data for each risk type respectively according to the determined risk type to be predicted and evaluating the data quality; making explicit risk judgments on the risk variable data based on a preset risk inference rule set; conducting implicit risk prediction on the risk variable data based on a pre-trained risk prediction model; evaluating the preliminary risks of each risk type based on explicit risks and implicit risks; calculating the explicit risk weights, implicit risk weights and preliminary risk weights of each risk type respectively based on data quality, explicit risks and implicit risks; calculating the comprehensive risk of each risk type respectively based on the explicit risk weights, implicit risk weights and preliminary risk weights.

2. The risk assessment method for the overall production plan of an electronic system as described in claim 1, wherein, The making explicit risk judgments on the risk variable data based on a preset risk inference rule set includes: calculating the rule coverage rates of the risk variable data respectively based on a preset risk inference rule set.

3. The risk assessment method for the overall production plan of the electronic system according to claim 2, characterized in that, The conducting implicit risk prediction on the risk variable data based on a pre-trained risk prediction model includes: conducting predictions on the risk variable data respectively based on a pre-trained risk prediction model to obtain the risk occurrence probabilities and confidence levels of each risk type.

4. The risk assessment method for the overall production plan of an electronic system according to claim 3, wherein, The evaluating the preliminary risks based on explicit risks and implicit risks includes: matching the explicit risks and implicit risks of each risk type based on the constructed knowledge base to obtain preliminary risk conclusions; the knowledge base contains judgment conditions and judgment conclusions for judging explicit risks and implicit risks.

5. The risk assessment method for the overall production plan of the electronic system according to claim 4, characterized in that The calculating the explicit risk weights, implicit risk weights and preliminary risk weights of each risk type respectively based on data quality, explicit risks and implicit risks includes: calculating the explicit risk weights, implicit risk weights and preliminary risk weights of each risk type respectively according to the following method: Explicit risk weight = 0.3 × rule coverage rate + 0.2 × data quality; Implicit risk weight = 0.4 × confidence level + 0.1 × data quality; Preliminary risk weight = 1 - explicit risk weight - implicit risk weight.

6. The risk assessment method for the overall production plan of an electronic system as described in claim 5, wherein The calculating the comprehensive risk of each risk type respectively based on the explicit risk weights, implicit risk weights and preliminary risk weights includes: calculating the comprehensive risk of each risk type respectively according to the following method: Comprehensive risk = (explicit risk contribution × explicit risk weight) + (risk occurrence probability × implicit risk weight) + (preliminary risk contribution × preliminary risk weight); wherein, the explicit risk contribution is obtained by quantifying the explicit risk, and the preliminary risk contribution is obtained by quantifying the preliminary risk conclusion.

7. The risk assessment method for the overall production plan of an electronic system according to any one of claims 1-6, characterized in that, It further includes: outputting the comprehensive risk of each risk type and the calculation process of each comprehensive risk.

8. An electronic system overall production plan risk assessment device, characterized in that including: a first unit for collecting risk variable data for each risk type respectively according to the determined risk type to be predicted and evaluating the data quality; a second unit for making explicit risk judgments on the risk variable data based on a preset risk inference rule set; a third unit for conducting implicit risk prediction on the risk variable data based on a pre-trained risk prediction model; a fourth unit for evaluating the preliminary risks of each risk type based on explicit risks and implicit risks; a fifth unit for calculating the explicit risk weights, implicit risk weights and preliminary risk weights of each risk type respectively based on data quality, explicit risks and implicit risks; The sixth unit is used to calculate the comprehensive risk of each risk type based on the explicit risk weight, the implicit risk weight, and the preliminary risk weight respectively.

9. An overall production plan risk assessment device for an electronic system, comprising a processor and a storage medium, characterized in that, The storage medium stores computer instructions, and the processor is connected to the storage medium and is used to run the computer instructions to execute the risk assessment method for the overall production plan of the electronic system according to any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is run by the processor, it can execute the risk assessment method for the overall production plan of the electronic system according to any one of claims 1-7.

Images