Electronic product quality evaluation comprehensive model design method

Abstract

The invention belongs to the technical field of electronic product quality evaluation, and discloses an electronic product quality evaluation comprehensive model design method. The quality evaluation comprehensive model design method comprises the steps of constructing a quality evaluation index system, inputting quality evaluation index system data corresponding to each stage, carrying out dimensionless processing by using an index value dimensionless method, carrying out weight processing by using an index weight method, and designing a quality evaluation comprehensive model. And performing comprehensive evaluation by using a quality comprehensive evaluation method to obtain a comprehensive evaluation result. According to the invention, electronic enterprises can be helped to better implement quality evaluation of complete machine-level, component-level and information system-level products, know the quality development trends of different electronic products in different stages, and further perform early warning on quality problems in different stages so as to reduce quality loss.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of electronic product quality evaluation, and in particular to a method for designing a comprehensive model for electronic product quality evaluation. Background Art

[0002] With the widespread application of information science and technology in the field of electronic products, electronic products are becoming more and more powerful and their structures are becoming more and more complex, which puts forward new and higher requirements for quality management. Traditional electronic product quality evaluation focuses on the quality evaluation of a single stage, and rarely involves the full life cycle evaluation of different types of electronic products such as the whole machine level, component level and information system level.

[0003] In order to use quality information resources scientifically, improve the utilization rate of quality information resources, and assist management departments in making decisions for quality level situation analysis, risk prediction, and other tasks in the future, it is very necessary to conduct a comprehensive assessment of the quality levels of electronic products during the development, production, and in-service stages. Using the assessment as a starting point, we can track the quality changes of electronic products in real time throughout the entire process and scientifically analyze the status of the entire system and its entire life cycle. Summary of the invention

[0004] The purpose of the present invention is to provide a method for designing a comprehensive model for electronic product quality evaluation. Based on a systematic study of the quality status and quality control requirements of electronic products, combined with product characteristics, an in-depth analysis of its quality information demand characteristics is conducted, and quality evaluation system indicators covering the stages of development, production, use and service are designed. The quality information of electronic products is classified and summarized, the correlation between them is clarified, and an electronic product quality information indicator system and a comprehensive quality evaluation model are constructed to objectively reflect the quality level of electronic products, further improve quality management efficiency, and provide strong support for improving product quality levels and improving decision-making.

[0005] To achieve the above object, the present invention provides a method for designing a comprehensive model for electronic product quality evaluation, comprising the following steps:

[0006] Step 1: For electronic products, a quality evaluation index system is established according to the quality requirements of the development stage, batch production stage and in-service stage;

[0007] Step 2: Input the quality evaluation index system data corresponding to each stage in step 1;

[0008] Step 3: For the quality evaluation index system data corresponding to each stage in step 1, use the index value dimensionless method to perform dimensionless processing;

[0009] Step 4: For the dimensionless processed data, use the indicator weight method to perform weight processing;

[0010] Step 5: For the data that has been weighted by the indicator weight method, use the comprehensive quality evaluation method to conduct a comprehensive evaluation and obtain the comprehensive evaluation results.

[0011] Furthermore, the indicator weight processing adopts the hierarchical analysis method to investigate the priority of qualitative indicators of quality evaluation, and let c1f c2fK fc n , based on this priority comparison, determine element a ij The value of c i fc j , then a ij >1; where c1fc2fK fc n Represents the importance of different indicators in quality evaluation. Experts determine the importance of each indicator in the quality evaluation process of electronic products. The higher the importance, the higher the priority;

[0012] Adjust the element aij with the highest priority of the qualitative indicator of violation evaluation, and its adjustment range is (aij-1, aij+1), that is, a i ' j Make sure its adjustment value is the scale value;

[0013] According to the interval weight requirement, let a certain indicator c i The weight interval is

[0014] Compare the judgment matrix with the interval weights of some indicators given in advance. If the weight obtained by the adjusted judgment matrix meets the interval weight requirements, the final weight is directly calculated; if the adjusted weight does not belong to the interval, adjust the second smallest element of the range indicator priority according to the interval weight requirements for the pairwise comparison elements of the judgment matrix until the consistency ratio of the judgment matrix is ​​met.

[0015] Furthermore, the quality evaluation system includes at least three levels of dimensions;

[0016] The first-level dimension of the development stage is development, the second-level dimension is the prototype stage and the design finalization stage, and the third-level dimension is the functional performance indicators, six quality indicators, and software corresponding to the prototype stage, and the design finalization, functional performance assessment indicators, general quality characteristic test assessment, software, test, and in-service assessment corresponding to the design finalization stage;

[0017] The first-level dimension of the batch production stage is batch production, the second-level dimension is the production stage and the use stage, and the third-level dimension is the product acceptance, product inspection, environmental routine test, reliability acceptance test, quality problems corresponding to the production stage, and the quality problems corresponding to the use stage;

[0018] The first-level dimension of the in-service stage is in-service, the second-level dimension is in-service, and the third-level dimension is quality issues.

[0019] Furthermore, the batch production stage in step 1 includes the production stage and the use stage from leaving the factory to formal delivery to the user unit.

[0020] Furthermore, the dimensionless methods for the index values ​​in step 3 include range transformation method, normalization method, standard deviation method, efficacy coefficient method, and linear proportional transformation method. In step 3, the range transformation method, normalization method, standard deviation method, efficacy coefficient method, and linear proportional transformation method are used in combination to cope with various dimensionless demand situations.

[0021] Furthermore, the indicator weighting method in step 4 includes subjective weighting method, objective weighting method and combined weighting method, and step 4 selects to use one of the subjective weighting method, objective weighting method and combined weighting method.

[0022] Furthermore, the comprehensive quality evaluation method in step 5 includes a comprehensive index method, an efficacy scoring method, a rank sum ratio method, and a regression fitting method. Step 5 selects to use one of the comprehensive index method, the efficacy scoring method, the rank sum ratio method, and the regression fitting method.

[0023] Beneficial effects:

[0024] The present invention provides a method for designing a comprehensive model for electronic product quality evaluation, and carries out quality evaluation work for electronic products throughout their life cycle. The main contents include aspects such as the design of an index system and the design of an evaluation index model. The electronic product quality information index system and the comprehensive quality evaluation model constructed by the present invention can objectively reflect the quality level of electronic products, and provide strong support for improving product quality and improving decision-making; the quality evaluation index provided by the present invention is oriented to its entire life cycle, the quality index design should fully consider the data source status, the weight design should comprehensively consider subjective and objective factors, and the evaluation model design adopts a linear synthesis method, which can more easily find influencing factors, thereby improving the quality level of electronic products; the present invention can help electronic companies better implement the quality evaluation of whole machine level, component level and information system level products, understand the quality development trends of different electronic products at different stages, and then warn of quality problems at different stages to reduce quality losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a process diagram of building a comprehensive model for electronic product quality evaluation involved in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The preferred mechanism and implementation method of the present invention are further described below in conjunction with the accompanying drawings and specific implementation methods.

[0027] like Figure 1As shown, the embodiment of the present invention discloses a technical solution for a method for designing a comprehensive model for evaluating the quality of electronic products. The goal of the method for designing a comprehensive model for evaluating the quality of electronic products provided by the present invention is to comprehensively and systematically grasp the overall quality operation status of electronic products, and to provide strong support for improving the quality management level and identifying the weak links in quality management. The quality evaluation of electronic products throughout their life cycle mainly focuses on solving the following problems, namely, the design of the quality evaluation index system, the design of the quality evaluation index weights, and the design of the quantitative quality evaluation model.

[0028] On the basis of reviewing the existing advanced evaluation models, a centralized and effective evaluation model is adopted and compatibility analysis is conducted to better understand the scientificity and effectiveness of the evaluation results. The specific evaluation index and model construction process is as follows: first, the advanced electronic product quality evaluation is studied and included in the database; second, the qualitative indicators are quantified, that is, the language scale design is carried out; and finally, the dimensionless index value method, the index weight method, and the comprehensive quality evaluation method are used to conduct empirical analysis and research on electronic products.

[0029] The specific evaluation indicators are defined as follows:

[0030] Dimensionless index value method: also known as data standardization or normalization, is a commonly used data preprocessing method. The main purpose of dimensionless is to eliminate the dimensional effects between different features or indicators through data transformation, so that the data is comparable for subsequent analysis. Currently, the commonly used dimensionless methods include range transformation method, normalization method, standard deviation method, power coefficient method, and linear proportional transformation method.

[0031] The range transformation method is a widely used data processing technology that can linearly transform the original data to form a new variable, which is more convenient for insight into the trend changes and feature correlations behind the data. This method plays an important role in machine learning tasks, especially in the current era of big data.

[0032] Normalization is a common technique in data preprocessing, which aims to scale the data so that it falls into a small specific interval, usually [0,1] or [-1,1]. This process can help improve the convergence speed and performance of the algorithm, especially when dealing with features of different dimensions or magnitudes.

[0033] The standard deviation is also called the standard deviation, which is the square root of the average of the squares of the data deviations from the mean. The standard deviation is the arithmetic square root of the variance, which can reflect the degree of dispersion of a data set. However, due to the square term of the variance, the multiple change of the dimension cannot directly reflect the degree of deviation, so the standard deviation appears. The smaller the standard deviation, the less these values ​​deviate from the mean, and vice versa.

[0034] Efficacy coefficient method: The efficacy coefficient method is also called the efficacy function method. It is based on the principle of multi-objective planning. It determines a satisfactory value and an unacceptable value for each evaluation indicator. The satisfactory value is the upper limit and the unacceptable value is the lower limit. The degree to which each indicator achieves the satisfactory value is calculated, and the score of each indicator is determined. The score is then combined through weighted average to evaluate the comprehensive situation of the research object.

[0035] Linear scaling method: It is a method of linearly transforming the original data. It is usually used to map the data to a specific range to eliminate the influence of dimension and order of magnitude and change the weight of the variable in the reanalysis.

[0036] Indicator weight method: Use questionnaire surveys, model analysis and other methods to divide indicators according to proportion and calculate the weight of each indicator. Currently, the commonly used indicator weight methods include subjective weighting method, objective weighting method and combined weighting method.

[0037] Subjective weighting method: It is a method to determine the weight of attributes, which determines the weight according to the decision maker's subjective emphasis on individual attributes. Common subjective weighting methods include expert survey method, hierarchical analysis method, binomial coefficient method, chain scoring method and least square method. Compared with the objective weighting method, the subjective weighting method has an advantage in determining the weight according to the decision maker's intention, but it is less objective and more subjective.

[0038] Objective weighting method: The weight of the indicator is determined according to the degree of variation of the values ​​of each indicator, avoiding the deviation caused by human factors. Compared with the subjective weighting method, it has higher accuracy and stronger objectivity, and can better explain the results.

[0039] Combined weighting method: It is a weighting method that combines subjective weighting method and objective weighting method to determine the weight of decision indicators. The specific steps include using the AHP method to determine the weight of each factor, using the output of the AHP method as the input of the entropy method to calculate the objective weight, and then combining the subjective weight and the objective weight to obtain the final weight value.

[0040] Currently, the commonly used comprehensive quality evaluation methods include comprehensive index method, efficacy scoring method, rank sum ratio method, and regression fitting method.

[0041] Comprehensive index method: It refers to a method of comprehensively evaluating economic benefits by taking the weighted average of the individual indexes of various economic indicators and calculating the comprehensive value of economic benefits on the basis of determining a reasonable economic benefit indicator system.

[0042] Rank sum ratio method: It is a statistical analysis method that combines the advantages of classical parametric statistics and modern non-parametric statistics. It is not only suitable for the comprehensive evaluation of four-cell table data, but also for the comprehensive evaluation of row × column table data, and also for the comprehensive evaluation of quantitative data and categorical data.

[0043] Example 1

[0044] Based on the quality control characteristics of electronic products, a quality evaluation index system was designed, which mainly includes three stages: research and development, batch production (production and use (from leaving the factory to before formal delivery to the user unit)) and in-service.

[0045] In this embodiment, a method for designing a comprehensive model for evaluating electronic product quality includes the following steps:

[0046] Step 1: For electronic products, a quality evaluation index system is established according to the quality requirements of the development stage, batch production stage and in-service stage;

[0047]

[0048]

[0049] Step 2: Enter the specific data related to the quality evaluation indicators at different stages; take the specific data from 2017 to 2019 as an example (where 1 means meeting the requirements, 42 / 155 means there are 42 defects in every 155 thousand lines of code, and 25 / 165 means there are 25 defects in every 165 thousand lines of code):

[0050]

[0051]

[0052] Step 3: For the quality evaluation index system data corresponding to each stage from 2017 to 2019, the index value dimensionless processing is performed using the index value dimensionless method (in the following table, 1 represents the best, 0 represents the worst, and the larger the value between the two, the better the index value and the greater the contribution to the quality level), and the result is:

[0053]

[0054]

[0055] Step 4: For the dimensionless processed data, use the indicator weight method to perform weight processing;

[0056] The indicator weight processing first adopts the hierarchical analysis method. The priority of the survey indicator, without loss of generality, is c1f c2f Kf c n , where c1 f c2f K fc n Represents the importance of different indicators in quality evaluation. Experts determine the importance of each indicator in the quality evaluation process of electronic products. The higher the importance, the higher the priority. ij If ci fc j , then a ij > 1. Following this principle, first adjust the element a that violates the highest priority index ij , its adjustment range is (aij-1, aij+1), that is, a i ' j Make sure that its adjustment value is a scale value (the benefit of gradual adjustment is mainly reflected in the gradual adjustment, which reduces the impact of changes in the judgment value of a single element on the entire judgment matrix).

[0057] If the adjusted element is a 12 , then the judgment matrix can be adjusted to: Calculate the consistency ratio of the adjusted matrix. If the consistency is met, the adjusted judgment matrix meets the requirements. If the consistency ratio of the adjusted matrix does not meet the consistency requirements, gradually adjust the next smallest element of the range indicator priority until the consistency ratio of the judgment matrix is ​​met. If the adjusted judgment matrix does not meet the priority requirements, improve it according to the requirements of the above steps until all priorities are met. Compare the judgment matrix with the interval weights of some indicators given in advance. If the weight obtained by the adjusted judgment matrix meets the interval weight requirements, directly calculate the final weight. If not, go directly to the next step. According to the interval weight requirements, let a certain indicator c i The weight interval is If the adjusted weight does not belong to this interval, Then, according to the interval weight requirements for the pairwise comparison elements of the judgment matrix, the element is adjusted, and under the consistency ratio constraint, some elements are fine-tuned with reference to the evaluation values ​​of other evaluators.

[0058] The weight results are shown in the following table, which shows the weight of the fourth-level dimension to the third-level dimension, and the final weight shows the weight of each fourth-level dimension in product quality evaluation. The sum of the final weights in the last column is 1.

[0060]

[0061]

[0062] Step 5: For the data that has been weighted by the indicator weight method, use the quality comprehensive evaluation method to conduct a comprehensive evaluation and obtain the comprehensive evaluation results. The comprehensive evaluation results are as follows:

[0063] Product Phase 2017 2018 2019 Development stage 0.2383 0.2383 0.2383 Batch production stage 0.3424 0.3464 0.3476 In-service stage 0.3865 0.3866 0.3855

[0064] In the above table, the maximum comprehensive evaluation values ​​of the development stage, batch production stage, and in-service stage are 0.3030, 0.3099, and 0.3872, respectively, and the minimum value is 0. The closer the sum of the three stages is to 1, the higher the overall quality level.

[0065] The present invention provides a method for designing a comprehensive model for electronic product quality evaluation, which can help electronic companies better implement quality evaluation of products at the whole machine level, component level and information system level, understand the quality development trends of different electronic products at different stages, and then provide early warning for quality problems at different stages to reduce quality losses.

[0066] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent substitutions for some of the technical features therein. However, any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for designing a comprehensive model for evaluating the quality of electronic products, characterized in that: The following steps are involved: Step 1: For electronic products, a quality evaluation index system is established according to the quality requirements of the development stage, batch production stage and in-service stage; Step 2: Input the quality evaluation index system data corresponding to each stage in step 1; Step 3: For the quality evaluation index system data corresponding to each stage in step 1, use the index value dimensionless method to perform dimensionless processing; Step 4: For the dimensionless processed data, use the indicator weight method to perform weight processing; Step 5: For the data that has been weighted by the indicator weight method, use the comprehensive quality evaluation method to conduct a comprehensive evaluation and obtain the comprehensive evaluation results.

2. The electronic product quality evaluation comprehensive model design method according to claim 1 is characterized in that: The indicator weight processing adopts the hierarchical analysis method to investigate the priority of qualitative indicators for quality evaluation, and let c1f c2fK fc n , based on this priority comparison, determine element a ij The value of c i fc j , then a ij >1; Adjust the element aij with the highest priority of violating the qualitative indicator of quality evaluation, and its adjustment range is (aij-1, aij+1), that is, a i ' j Make sure its adjustment value is the scale value; According to the interval weight requirement, let a certain indicator c i The weight interval is Compare the judgment matrix with the interval weights of some indicators given in advance. If the weights obtained by the adjusted judgment matrix meet the interval weight requirements, directly calculate the final weights; If the adjusted weight does not belong to the interval, the next smallest element of the range indicator priority is adjusted according to the interval weight requirements for the pairwise comparison elements of the judgment matrix until the consistency ratio of the judgment matrix is ​​met.

3. The electronic product quality evaluation comprehensive model design method according to claim 1 is characterized in that: The quality evaluation system includes at least three levels of dimensions; The first-level dimension of the development stage is development, the second-level dimension is the prototype stage and the design finalization stage, and the third-level dimension is the functional performance indicators, six quality indicators, and software corresponding to the prototype stage, and the design finalization, functional performance assessment indicators, general quality characteristic test assessment, software, test, and in-service assessment corresponding to the design finalization stage; The first-level dimension of the batch production stage is batch production, the second-level dimension is the production stage and the use stage, and the third-level dimension is the product acceptance, product inspection, environmental routine test, reliability acceptance test, quality problems corresponding to the production stage, and the quality problems corresponding to the use stage; The first-level dimension of the in-service stage is in-service, the second-level dimension is in-service, and the third-level dimension is quality issues.

4. The electronic product quality evaluation comprehensive model design method according to claim 1 is characterized in that: The batch production stage in step 1 includes the production stage and the use stage after leaving the factory until it is officially delivered to the user unit.

5. The electronic product quality evaluation comprehensive model design method according to claim 1 is characterized in that: The methods for dimensionlessizing the index values ​​in step 3 include range transformation method, normalization method, standard deviation method, efficacy coefficient method, and linear proportional transformation method. In step 3, the range transformation method, normalization method, standard deviation method, efficacy coefficient method, and linear proportional transformation method are used in combination.

6. The electronic product quality evaluation comprehensive model design method according to claim 1 is characterized in that: The indicator weighting methods in step 4 include subjective weighting method, objective weighting method and combined weighting method. Step 4 selects to use one of the subjective weighting method, objective weighting method and combined weighting method.

7. The electronic product quality evaluation comprehensive model design method according to claim 1 is characterized in that: The comprehensive quality evaluation methods in step 5 include comprehensive index method, efficacy scoring method, rank sum ratio method, and regression fitting method. Step 5 selects to use one of the comprehensive index method, efficacy scoring method, rank sum ratio method, and regression fitting method.

Images