CN114610274A - Improved demand priority evaluation method based on analytic hierarchy process - Google Patents

Abstract

The invention relates to an improved requirement priority evaluation method based on an analytic hierarchy process, and belongs to the field of software engineering. Aiming at a software system construction related party, determining N items of evaluation factors by an expert judgment method, and establishing a hierarchical structure model; then, constructing a judgment matrix based on the determined evaluation factors, calculating the weight coefficients of the evaluation factors by using a sum-product method, and carrying out consistency check; finally, respectively to the functional modules of the software system from F₁、F₂，……，F_NAnd scoring the N evaluation dimensions, and performing weighted calculation to obtain a priority sequence of the functional module. In the technical scheme provided by the invention, the evaluation factors are determined by using an expert judgment method, so that the evaluation factors are more in line with the actual situation. The method provided by the invention can be used for calculating the priority of the software requirements more scientifically and reasonably and evaluating the development sequence of the software requirements.

Description

An Improved Requirement Priority Evaluation Method Based on Analytic Hierarchy Process

technical field

The invention belongs to the field of software engineering, and in particular relates to an improved method for evaluating requirements priority based on AHP.

Background technique

With the vigorous development of computer technology, software systems are applied in all aspects of life and work. In the process of designing and developing a software system, how to obtain the requirement priority accurately and quickly is of great significance to the software system itself and its development enterprises.

The assessment methods of demand priority are mainly divided into two categories: one is a qualitative method that evaluates needs subjectively based on experience; the other is a quantitative method that evaluates based on survey data. At present, the widely used requirements priority assessment methods include the Kano model and the AHP.

The Kano model divides needs into five types: basic needs, expected needs, exciting needs, indifference needs, and reverse needs. The model establishes a two-dimensional model to measure the user's satisfaction with the functional characteristics of the product. By asking questions about the same function point in both forward (strengthened) and reverse (weakened) aspects, the user's subjective feelings are obtained, and the user's answer The mutual exclusion relationship ensures the authenticity and validity of the user's answer results. Finally, the questionnaire results are classified and summarized according to the five types of needs, and a two-dimensional "satisfaction influence" is established as the X-axis and "unsatisfactory influence" as the Y-axis. Models that form quantifiable requirements development priorities. When the Kano model is applied to the research of new functions, these functions must not be available in their own products, and must be functions that other products have or are easily understood by users through descriptions. The model cannot be applied to innovative functions that are inconvenient to describe. .

Analytic Hierarchy Process is a comprehensive evaluation method for systematic analysis and decision-making, and it is a process of quantifying qualitative problems. Its main feature is that through the establishment of a hierarchical structure, the judgment is transformed into several factors, and the importance of the two is compared, so that the qualitative judgment that is difficult to quantify is transformed into an operable comparison of the importance. Its essence is to decompose complex problems into multiple components, and then form these factors into a hierarchical structure according to the dominance relationship, and determine the ranking through the method of pairwise comparison, which overcomes the shortcomings of other methods that avoid the subjective judgment of decision makers. This method is not suitable for too many decision-making layers, and requires that the indicator data in the decision-making be known.

SUMMARY OF THE INVENTION

(1) Technical problems to be solved

The technical problem to be solved by the present invention is how to provide an improved requirement priority assessment method based on AHP, so as to solve the problem of accurately and rapidly obtaining the requirement priority during the design and development process of the software system.

(2) Technical solutions

In order to solve the above-mentioned technical problems, the present invention proposes an improved method for evaluating requirements priority based on AHP, and the method includes the following steps:

S1. Determine the evaluation factors: For the parties involved in the construction of the software system, determine the N items of evaluation factors through the expert judgment method;

S2. Establish a hierarchical structure model: The hierarchical structure model includes a scheme layer, a criterion layer and a target layer. The scheme layer is the functional module of the software system, the criterion layer is the selected evaluation factors and corresponding weights, and the target layer is the software system function module. level sequence;

S3. Construct judgment matrix: construct an N-order judgment matrix based on the determined N evaluation factors;

S4. Calculate the weight coefficient and consistency index: use the sum-product method to analyze the N-order judgment matrix, calculate the weight coefficient of the evaluation factors, and calculate the judgment matrix consistency index CI;

S5, consistency check: according to the judgment matrix consistency index CI, judge whether the consistency result is passed, if not, then it is necessary to return to step S3 to adjust the judgment matrix, and perform weight calculation again; if passed, then execute step S6;

S6. Calculate weighted score: Score each functional module of the software system from N evaluation dimensions of F ₁ , F ₂ , ... , F _N , and perform weighted calculation to obtain the priority sequence of the functional modules.

Further, the step S1 specifically includes: based on the relevant parties of the software system construction, determining that the scale of the personnel participating in the evaluation factor is M; listing the factors X items that affect the software priority; and the participants in the software system construction according to the influence determined by themselves. According to the size of the degree, _{N items of} influencing factors are selected, and the votes of the influencing factors are counted according to the selection of participants.

Further, the relevant parties of the software system construction include project leaders, system architecture designers, developers and testers.

Further, the scale of the personnel participating in the evaluation factor is determined according to the complexity of the software system.

Further, the step S3 specifically includes:

The participants in the software system construction compare the evaluation factors F ₁ , F ₂ , ..., F _N in pairs to form a judgment matrix P(F _i , F _j )', i,j=1,2,..., N, where P(F _i , F _j ) represents the ratio of the importance of factor F _i to factor F _j , called scale, P(F _i , F _j )=1/P(F _j , F _i ) .

Further, in the step S4, the sum-product method is used to analyze the N-order judgment matrix, and the calculation of the weight coefficient of the evaluation factor specifically includes:

S41, normalize each column element of the judgment matrix

S42, the judgment matrix after the normalization process is completed is added by columns

S43, normalize the vector W'=(w' ₁ , w' ₂ , ..., w' _N ) ^T

The approximate value W=(w ₁ , w ₂ , ..., w _N ) of the feature vector is obtained, and w ₁ , w ₂ , ..., w _N is the weight of the evaluation factors F ₁ , F ₂ , ... , F _N .

Further, calculating the judgment matrix consistency index CI in the step S4 specifically includes:

S44. Calculate the maximum eigenvalue λ _max of the judgment matrix

S45. Calculate the Consistency Index CI (Consistency Index) of the judgment matrix

当CR<阈值时，一致性结果为通过，否则不通过。Further, in the step S5, judging whether the consistency result is passed according to the judgment matrix consistency index CI specifically includes: searching the average random consistency index RI (RandomIndex) from the random consistency table according to the number of evaluation factors N, and then calculating the random consistency index RI (RandomIndex). Consistency ratio CR (ConsistencyRatio),

When CR<threshold, the consistency result is passed, otherwise it is not passed.

Further, the threshold=0.1.

Further, the step S6 specifically includes: respectively grading the functional modules of the software system from N evaluation dimensions of F ₁ , F ₂ , ..., F _N , using F ₁ , F ₂ , ... , F _N The _weights w ₁ , w ₂ , .

(3) Beneficial effects

The present invention proposes an improved demand priority assessment method based on AHP. Compared with the existing demand priority assessment method, the technical solution proposed by the present invention first uses the expert judgment method to determine the assessment factors, so that the assessment factors are more in line with the actual situation. With the method proposed in the present invention, the priority of software requirements is calculated more scientifically and reasonably, and the development sequence of software requirements is evaluated.

Description of drawings

Fig. 1 is the flow chart of the processing procedure method of the present invention;

Fig. 2 is the present invention to establish the hierarchical structure model structure diagram;

Fig. 3 is the structure diagram of the hierarchical structure model established based on the application asset system.

Detailed ways

In order to make the purpose, content and advantages of the present invention clearer, the specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments.

The purpose of the present invention is to propose a requirement priority assessment method, which utilizes the improved AHP to assess the software requirement priority.

1 Overall plan

The specific process of the improved AHP-based demand priority evaluation scheme proposed by the present invention is shown in Figure 1, and the specific process is: Personnel, testers, etc., determine N items of evaluation factors through expert judgment (the number of evaluation factors can be determined according to the complexity of the software system), and establish a hierarchical structure model; then, based on the determined N evaluation factors, construct an N-order judgment matrix, The weight coefficients of the evaluation factors are calculated by the sum-product method, and the consistency test is carried out; finally, each functional module of the software system is scored from the N evaluation dimensions of F ₁ , F ₂ , ... , F _N and weighted Calculate to get the priority sequence of function modules.

The detailed steps are as follows:

S1. Determine the evaluation factors

Based on the relevant parties of the software system construction, such as project leaders, system architecture designers, developers, testers, etc., determine the scale of personnel involved in the evaluation factors as M (according to the complexity of the software system, determine the scale of personnel involved); Usually, X items of factors that affect the software priority are found; the participants in the software system construction select N items of influencing factors according to the degree of influence determined by themselves (generally, the number of influencing factors that need to be selected for the project is determined), and then participate in the The selection of personnel shall be counted for the influencing factors, and the N items with the largest number are determined as the evaluation factors F ₁ , F ₂ , ..., F _N .

S2, establish a hierarchical structure model

The hierarchical structure model is shown in Figure 2, including the scheme layer, the criterion layer and the target layer. The scheme layer is the software system function modules: function module 1, function module 2... function module n, and the criterion layer is the selected evaluation factors and Corresponding to the weight, the target layer is the priority sequence of software system function modules.

S3, construct judgment matrix

The participants in the construction of the software system compare the evaluation factors F ₁ , F ₂ , ..., F _N in pairs to form the following judgment matrix P(F _i , F _j )', i,j=1,2,... ,N, where P(F _i ,F _j ) represents the ratio of the importance of factor F _i to factor F _j , called scale, P(F _i ,F _j )=1/P(F _j ,F _i ). Wherein P(F ₁ , F ₂ ) represents the ratio of the importance of factor F ₁ to factor F ₂ , which is called scale, P(F ₁ , F ₂ )=1/P(F ₂ , F ₁ ).

Table 1 Judgment Matrix

F₁ F₂ ... F_N F₁ 1 P(F₂,F₁) ... P(F_N,F₁) F₂ P(F₁,F₂) 1 ... P(F_N,F₂) ... ... ... 1 ... F_N P(F₁,F_N) P(F₂,F_N) ... 1

The scale is the assignment of importance, and the meaning and value are shown in Table 2.

Table 2 Scale table

S4, calculate the weight coefficient and consistency index

The N-order judgment matrix constructed by the determined N evaluation factors is analyzed by the sum-product method. The specific calculation steps are as follows:

S41, normalize each column element of the judgment matrix

S42, the judgment matrix after the normalization process is completed is added by columns

S43, normalize the vector W'=(w' ₁ , w' ₂ , ..., w' _N ) ^T

Obtain the approximate value of the feature vector W=(w ₁ , w ₂ ,..., w _N ), w ₁ , w ₂ ,..., w _N is the weight of the evaluation factors F ₁ , F ₂ ,..., F _N

S44. Calculate the maximum eigenvalue λ _max of the judgment matrix

S45. Calculate the Consistency Index CI (Consistency Index) of the judgment matrix

S5. Consistency check

阈值＝0.1，当CR<阈值时，一致性结果为通过，否则不通过。如果不通过，则需要调整判断矩阵，重新进行权重计算。The average random consistency index RI (Random Index) is searched from the random consistency table (see Table 3) according to the number N of evaluation factors. Table 3 shows the RI values of orders 1 to 15. Then calculate the random consistency ratio CR (Consistency Ratio),

Threshold=0.1, when CR<threshold, the consistency result is passed, otherwise it is not passed. If it does not pass, you need to adjust the judgment matrix and recalculate the weights.

Table 3 Random Consistency Table

N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 RI 0 0 0.58 0.90 1.12 1.24 1.32 1.41 1.46 1.49 1.52 1.54 1.56 1.58 1.59

S6. Calculate the weighted score

The functional modules of the software system are respectively scored from the N evaluation dimensions of F ₁ , F ₂ , ..., F _N . _Using the _{weights w 1 , w 2 ,} .

2 Application examples

The following takes the application asset system as an example for specific description.

(1) Establish a hierarchical structure model based on the application asset system

First, the evaluation factors are determined by the expert judgment method. The scale of personnel involved in the evaluation factors is 20, and 10 factors that affect the software priority are selected. The final evaluation factors of scale, complexity, degree of reuse, and benefit expectation are determined by expert judgment. .

Establish a hierarchical structure model, Figure 3 is the hierarchical structure model of the application asset system established based on the target layer, the standard layer, and the solution layer.

(2) Constructing judgment matrix

The scale, complexity, reuse, and benefit expectations of the evaluation factors are compared in pairs, and the following judgment matrix is formed:

Table 4 Application Assets System Judgment Matrix

scale Complexity degree of reuse Expected benefit scale 1 0.5 0.25 2.5 Complexity 2 1 0.5 2 degree of reuse 4 2 1 4 Expected benefits 0.4 0.5 0.25 1

(3) Calculate the weight coefficient

Using the sum-product method to analyze the fourth-order judgment matrix constructed by the four evaluation factors of scale, complexity, degree of reuse, and expected benefit, the analysis obtains the eigenvector λ _max = (0.648, 0.981, 1.962, 0.409), the corresponding weight value They are: 16.207%, 24.520%, 49.040%, 10.233%. In addition, combined with the eigenvectors, the maximum eigenvalue can be calculated to be 4.109, and then the CI value of 0.036 can be calculated by using the maximum eigenvalue. The analysis results are shown in the following table:

Table 5 Analysis results

(4) Consistency test

Find the RI value of 0.90 from the random consistency RI table, then calculate the consistency index CR value (CR=CI/RI), and calculate that CR=0.041<threshold value, and the consistency result is passed. This result means that the judgment matrix in this study satisfies the consistency test, and the calculated weights are consistent.

(5) Calculate the weighted score

The four evaluation dimensions of the asset warehousing module, asset warehousing module, QR code generation module, and asset statistics query module of the application asset system are respectively scored in terms of scale, complexity, degree of reuse, and expected benefit. The scale, complexity, degree of reuse, and expected benefit are weighted to calculate the score of the functional module. The higher the score, the higher the priority, and the priority sequence can be obtained.

The above are only the preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, without departing from the technical principles of the present invention, several improvements and modifications can be made. These improvements and modifications It should also be regarded as the protection scope of the present invention.

Claims (10)

1. An improved method for demand priority assessment based on an analytic hierarchy process, the method comprising the steps of:

s1, determining evaluation factors: determining N evaluation factors by an expert judgment method aiming at software system construction related parties;

s2, establishing a hierarchical structure model: the hierarchical structure model comprises a scheme layer, a criterion layer and a target layer, wherein the scheme layer is a functional module of the software system, the criterion layer is selected evaluation factors and corresponding weights, and the target layer is a priority sequence of the functional module of the software system;

s3, constructing a judgment matrix: constructing an N-order judgment matrix based on the determined N evaluation factors;

s4, calculating a weight coefficient and a consistency index: analyzing the N-order judgment matrix by using a sum-product method, calculating a weight coefficient of an evaluation factor, and calculating a judgment matrix consistency index CI;

s5, consistency check: judging whether the consistency result passes according to the judgment matrix consistency index CI, if not, returning to the step S3 to adjust the judgment matrix, and carrying out weight calculation again; if so, executing step S6;

s6, calculating a weighted score: for each function module of software system from F₁、F₂，……，F_NAnd scoring the N evaluation dimensions, and performing weighted calculation to obtain a priority sequence of the functional module.

2. The improved analytic hierarchy-based demand priority assessment method of claim 1, wherein the step S1 specifically comprises: establishing related parties based on a software system, and determining the scale of people participating in evaluation factors as M people; listing factors X influencing the software priority; the participators of the software system construction select N influencing factors according to the influence degree judged by the participators, the influence factor is counted according to the selection condition of the participators, and the N items with the largest quantity are determined as the evaluation factors F₁、F₂，……，F_N。

3. The improved analytic hierarchy-based demand priority assessment method of claim 2, wherein the software system construction related parties comprise project managers, system architecture designers, developers, and testers.

4. The improved analytic hierarchy-based demand priority assessment method of claim 2, wherein the size of the personnel participating in the assessment factor is determined by the complexity of the software system.

5. The improved analytic hierarchy-based demand priority assessment method of any one of claims 1 to 4, wherein the step S3 specifically comprises:

evaluation factor F for participant constructed by software system₁、F₂，……，F_NComparing every two to form a judgment matrix P (F)_i,F_j) ', i, j ═ 1,2, … …, N, where P (F)_i,F_j) Representing factor F_iRelative to factor F_jIs called the scale, P (F)_i,F_j)＝1/P(F_j,F_i)。

6. The improved analytic hierarchy process based demand priority evaluation method of claim 5, wherein the step S4 is performed by analyzing an N-th order decision matrix using a sum-product method, and calculating the weight coefficient of the evaluation factor specifically comprises:

s41, normalizing each row of elements of the judgment matrix

S42, adding the judgment matrixes after the normalization processing according to columns

S43, pair vector W ═ W'₁,w'₂,……，w'_N)^TPerforming normalization processing

Obtain an approximate value W ═ W of the eigenvector₁,w₂,……，w_N)，w₁,w₂,……，w_NIs the evaluation factor F₁、F₂，……，F_NThe weight of (c).

7. The improved analytic hierarchy process based demand priority assessment method of claim 6, wherein the step S4 of calculating the judgment matrix consistency index CI specifically comprises:

s44, calculating and judging the maximum eigenvalue lambda of the matrix_max

S45, calculating and judging matrix consistency index CI (consistency index)

8. The improved analytic hierarchy process based demand priority evaluation method of claim 7, wherein the step S5 of judging whether the consistency result passes the specific steps according to the judgment matrix consistency index CI comprises: searching average random consistency index RI (random index) from the random consistency table according to the number N of the evaluation factors, then calculating a random consistency ratio CR (consistency ratio),

when CR is reached<When the threshold value is reached, the consistency result is passed, otherwise, the consistency result is not passed.

9. The improved analytic hierarchy-based demand priority assessment method of claim 8, wherein the threshold is 0.1.

10. The improved analytic hierarchy-based demand priority assessment method of claim 8 or 9, whereinThe step S6 specifically includes: from F to the functional modules of the software system respectively₁、F₂，……，F_NThe N evaluation dimensions are scored using F₁、F₂，……，F_NWeight w of₁,w₂,……，w_NAnd (4) calculating the score of the functional module by weighting, wherein the priority with high score is high, namely the priority sequence of the functional module is obtained.

Images