CN102447593A - Test method and device - Google Patents

Abstract

The invention discloses a test method and a test device, which are used for solving the problem that in the prior art the correctness for testing network equipment is low. The method comprises the following steps of testing each index of the network equipment for many times, obtaining multiple test values corresponding to each index; structuring a fuzzy evaluation matrix according to relative importance of degree of value between indexes; determining a weighted set according to the structured fuzzy evaluation matrix, and determining the percentage of each tested index on each performance level, thereby determining the evaluation matrix; multiplying the determined weighted set and the evaluation matrix, obtaining the performance evaluation set of the network equipment, and determining the performance level corresponding to the maximum element as the obtained test result. Through the method provided by the invention, each index is tested by the test device, and the influence of each index to the comprehensive performance of the network equipment is integrated, so that the comprehensive performance of the network equipment can be tested correctly.

Description

A kind of method of testing and device

Technical field

The present invention relates to communication technical field, relate in particular to a kind of method of testing and device.

Background technology

In communication network, exist the various network equipments, the quality of network equipment combination property has often determined the quality of communication network.In order to select the reasonable network equipment of combination property for use,, just need test each alternative network equipment to improve the quality of communication network.

With the router is that example describes.The index that influences the router combination property exists a variety of, and for example indexs such as the throughput of router, time delay can influence the combination property of router.

The method of in the prior art router being tested is: to some or certain several index, router is tested, the user is according to the test data that obtains, the quality of artificial judgement router combination property.

Yet each index of router more or less all can influence the combination property of router, only tests the inevitable combination property that can accurately not reflect router to some or certain several index.And the index that each user is stressed when judging router combination property good and bad also varies with each individual, and for example user 1 lays particular emphasis on the throughput of router; If the throughput performance of router is better; Think that then the combination property of router is better, and user 2 lays particular emphasis on the time delay of router, if the delay performance of router is better; Think that then the combination property of router is better; But no matter being throughput or time delay, all is the factor that influences the router combination property, only lays particular emphasis on the combination property that the resulting test result of some indexs can not reflect router accurately.Therefore, the accuracy to network equipment detection is lower in the prior art.

Summary of the invention

The embodiment of the invention provides a kind of method of testing and device, in order to solve the problem low to the accuracy of network equipment detection in the prior art.

A kind of method of testing that the embodiment of the invention provides comprises:

Each index to the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each index;

Based on the corresponding a plurality of test values of each index that obtain; Confirm the relative importance degree value between each index; With the relative importance degree value between each index of confirming is element structure fuzzy evaluation matrix, confirms weight sets based on the fuzzy evaluation matrix of structure;

To each index; Based on the corresponding a plurality of test values of this index that obtain; And the corresponding index test value scope of setting of each performance rate; Confirming the percentage of this index on each performance rate of test, is that the element structure is passed judgment on subclass with the percentage of this index on each performance rate of testing;

According to the judge subclass of confirming to each index respectively; Confirm to pass judgment on matrix; With weight sets of confirming and judge matrix multiple, the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on the corresponding performance rate of element of concentrating maximum confirm as the test result that obtains.

A kind of method of testing that the embodiment of the invention provides comprises:

Each two-level index to the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each two-level index;

Each first class index to the said network equipment; Based on the corresponding a plurality of test values of each two-level index under this first class index that obtains; Confirm the relative significance level value between each two-level index under this first class index; With the relative significance level value between each two-level index of confirming is the corresponding secondary fuzzy evaluation matrix of this first class index of element structure; Based on the secondary fuzzy evaluation matrix of structure, confirm the secondary weight sets that this first class index is corresponding;

To each two-level index under this first class index; Based on the corresponding a plurality of test values of this two-level index that obtain; And the corresponding two-level index test value scope of setting of each performance rate; Confirming the percentage of this two-level index on each performance rate of test, is that the corresponding secondary of this two-level index of element structure is passed judgment on subclass with the percentage of this two-level index on each performance rate of testing;

According to passing judgment on subclass to the secondary of each the two-level index structure under this first class index respectively; Confirm the secondary judge matrix that this first class index is corresponding; The secondary that the secondary weight sets that this first class index is corresponding is corresponding with this first class index is passed judgment on matrix multiple, obtains the corresponding one-level of this first class index and passes judgment on collection;

One-level according to confirming to each first class index of the said network equipment is respectively passed judgment on collection, confirms one-level judge matrix;

Confirm the relative significance level value between each first class index of the said network equipment; With the relative significance level value between each first class index is element structure one-level fuzzy evaluation matrix; One-level fuzzy evaluation matrix according to structure is confirmed the one-level weight sets; One-level weight sets of confirming and one-level are passed judgment on matrix multiple, and the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

A kind of testing apparatus that the embodiment of the invention provides comprises:

Test module is used for each index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each index;

The weight sets determination module; Be used for according to the corresponding a plurality of test values of each index that obtain; Confirm the relative importance degree value between each index; With the relative importance degree value between each index of confirming is element structure fuzzy evaluation matrix, confirms weight sets according to the fuzzy evaluation matrix of structure;

Pass judgment on the subclass determination module; Be used for to each index; Based on the corresponding a plurality of test values of this index that obtain; And the corresponding index test value scope of setting of each performance rate; Confirming the percentage of this index on each performance rate of test, is that the element structure is passed judgment on subclass with the percentage of this index on each performance rate of testing;

The test result determination module; Be used for according to being directed against the judge subclass that each index is confirmed respectively; Confirm to pass judgment on matrix; With weight sets of confirming and judge matrix multiple, the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on the corresponding performance rate of element of concentrating maximum confirm as the test result that obtains.

A kind of testing apparatus that the embodiment of the invention provides comprises:

Test module is used for each two-level index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each two-level index;

Secondary weight sets determination module; Be used for each first class index to the said network equipment; According to the corresponding a plurality of test values of each two-level index under this first class index that obtains, confirm the relative significance level value between each two-level index under this first class index, be the corresponding secondary fuzzy evaluation matrix of this first class index of element structure with the relative significance level value between each two-level index of confirming; According to the secondary fuzzy evaluation matrix of structure, confirm the secondary weight sets that this first class index is corresponding;

Secondary is passed judgment on the subclass determination module; Be used for to each two-level index under this first class index; According to this two-level index that obtains corresponding to individual test value; And the corresponding two-level index test value scope of setting of each performance rate; Confirming the percentage of this two-level index on each performance rate of test, is that the corresponding secondary of this two-level index of element structure is passed judgment on subclass with the percentage of this two-level index on each performance rate of testing;

One-level is passed judgment on the collection determination module; Be used for according to passing judgment on subclass to the secondary of each the two-level index structure under this first class index respectively; Confirm the secondary judge matrix that this first class index is corresponding; The secondary that the secondary weight sets that this first class index is corresponding is corresponding with this first class index is passed judgment on matrix multiple, obtains the corresponding one-level of this first class index and passes judgment on collection;

One-level judge matrix determination module is used for passing judgment on collection based on the one-level of confirming to each first class index of the said network equipment respectively, confirms one-level judge matrix;

The test result determination module; Be used for confirming the relative significance level value between each first class index of the said network equipment; With the relative significance level value between each first class index is element structure one-level fuzzy evaluation matrix; One-level fuzzy evaluation matrix according to structure is confirmed the one-level weight sets; One-level weight sets of confirming and one-level are passed judgment on matrix multiple, and the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

The embodiment of the invention provides a kind of method of testing and device; This method is repeatedly tested each index of the network equipment, obtains the corresponding a plurality of test values of each index, according to the structure of the relative importance degree value between each index fuzzy evaluation matrix; Fuzzy evaluation matrix according to structure is confirmed weight sets; And the definite percentage of each index on each performance rate of testing, confirm to pass judgment on matrix in view of the above, with weight sets of confirming and judge matrix multiple; The performance that obtains this network equipment is passed judgment on collection, and the wherein maximum corresponding performance rate of element is confirmed as the test result that obtains.Through said method, testing apparatus is all tested each index, and combines the influence of each index to the combination property of the network equipment, therefore can test accurately the combination property of the network equipment.

Description of drawings

The test process that Fig. 1 provides for the embodiment of the invention;

The another kind of test process that Fig. 2 provides for the embodiment of the invention;

Fig. 3 testing apparatus structural representation corresponding that provide for the embodiment of the invention with method of testing shown in Figure 1;

Fig. 4 testing apparatus structural representation corresponding that provide for the embodiment of the invention with another kind of method of testing shown in Figure 2.

Embodiment

Because each index of the network equipment more or less all can influence the combination property of the network equipment; Therefore only according to test value to some or certain several index; And the quality through artificial judgement network equipment combination property, resulting test result must be inaccurate, selects the network equipment that uses according to such test result; The combination property that the network equipment that can cause selecting shows in practical application is also not fully up to expectations, and network quality is descended.The embodiment of the invention provides a kind of method of testing; This method testing apparatus is repeatedly tested each index of the network equipment; And comprehensively each index draws test result to the influence of network equipment combination property with this, can test accurately the combination property of the network equipment.

Below in conjunction with Figure of description, the embodiment of the invention is described in detail.

Fig. 1 is the test process that the embodiment of the invention provides, and specifically may further comprise the steps:

S101: each index to the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each index.

In embodiments of the present invention, testing apparatus has all carried out repeatedly test to each index of the network equipment, and the number of times of concrete test can be set for example 100 times as required.

For example, each index of the network equipment is as shown in table 1:

Table 1

As shown in table 1; The index of this network equipment comprises: index 1, index 2, index 3, index 4; These four indexs all can influence the overall target of this network equipment, and therefore, testing apparatus is repeatedly tested respectively these four indexs; For example respectively these four indexs are carried out 100 tests, obtain 100 corresponding respectively test values of these four indexs.

S102:, confirm the relative importance degree value between each index according to the corresponding a plurality of test values of each index that obtain.

In embodiments of the present invention; Testing apparatus can be according to the test value that obtains; Confirm the relative importance degree value between each index; The relative importance degree value is represented: each index of this network equipment is compared in twos, and the degree that one of them index influences network equipment combination property influences the comparison value of the degree of network equipment combination property than another index.

S103: with the relative importance degree value between each index of confirming is element structure fuzzy evaluation matrix, confirms weight sets according to the fuzzy evaluation matrix of structure.

After testing apparatus has been confirmed the relative importance degree value between each index; Be element structure fuzzy evaluation matrix then with each relative importance degree value; And definite in view of the above weight sets; Each element representation in this weight sets: after having taken all factors into consideration the relative importance degree value between each index, each index influences the degree value of network equipment combination property accordingly.

S104: to each index; According to the corresponding a plurality of test values of this index that obtain; And the corresponding index test value scope of setting of each performance rate; Confirming the percentage of this index on each performance rate of test, is that the element structure is passed judgment on subclass with the percentage of this index on each performance rate of testing.

In embodiments of the present invention, can preestablish each performance rate, for example be set at four performance rates, be respectively: the first estate, second grade, the tertiary gradient, the fourth estate.And can be directed against each index, set the corresponding index test value scope of each performance rate.Testing apparatus then can be directed against each index; Confirm in the corresponding test value of this index; Drop on the quantity of the test value in the corresponding index test value scope of each performance rate respectively, and then confirm the percentage of this index on each performance rate, and structure is passed judgment on subclass in view of the above.

S105:, confirm to pass judgment on matrix according to the judge subclass of confirming to each index respectively.

In embodiments of the present invention, confirmed the judge subclass of each index respectively after, then confirm to pass judgment on matrix with each element of passing judgment in subclass, comprised in this judge matrix and respectively passed judgment on all elements in the subclass.

S106: with weight sets of confirming and judge matrix multiple, the performance that obtains this network equipment is passed judgment on collection, this performance is passed judgment on the corresponding performance rate of element of concentrating maximum confirm as the test result that obtains.

Because each index of each element representation in the weight sets influences the degree value of network equipment combination property; Pass judgment on the percentage of certain index of each element representation on certain performance rate in the matrix; Therefore with weight sets and judge matrix multiple; Resulting performance is passed judgment on the percentage of combination property on each performance rate that each element of concentrating is the network equipment, therefore performance is passed judgment on and is concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

In said process; Testing apparatus is repeatedly tested each index of the network equipment, obtains the corresponding a plurality of test values of each index, according to the structure of the relative importance degree value between each index fuzzy evaluation matrix; Fuzzy evaluation matrix according to structure is confirmed weight sets; And the definite percentage of each index on each performance rate of testing, confirm to pass judgment on matrix in view of the above, with weight sets of confirming and judge matrix multiple; The performance that obtains this network equipment is passed judgment on collection, and the wherein maximum corresponding performance rate of element is confirmed as the test result that obtains.Through said method, testing apparatus is all tested each index, and combines the influence of each index to the combination property of the network equipment, therefore can test accurately the combination property of the network equipment.

In step S101 shown in Figure 1, testing apparatus carries out the test of set point number according to set point number to each index of the network equipment, obtains the corresponding a plurality of test values of each index.Wherein, to each index, the quantity of a plurality of test values that this index that obtains is corresponding equates with this set point number.And in order further to improve the accuracy of test, it is bigger that this set point number can be set, preferable, and this set point number is not less than the number of the index of the network equipment.

In step S102 shown in Figure 1; Testing apparatus confirms that according to the corresponding a plurality of test values of each index that obtain the method for the relative importance degree value between each index is: the corresponding a plurality of test values of each index to obtain are that element is constructed first test matrix:

$A={\left(\begin{array}{cccc}{x}_{11}& x_{12}& \·\·\·& x_{1m}\\ x_{21}& x_{22}& \·\·\·& x_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ x_{n1}& x_{n2}& \·\·\·& x_{\mathrm{nm}}\end{array}\right)}_{n\×m},$

Wherein, n is said set point number, and m is the number of the index of the said network equipment, x _IjBe the i time resulting test value of test that j index in m the index carried out; First test matrix according to structure is confirmed the relative importance degree value between each index.

When confirming the relative importance degree value between each index according to above-mentioned first test matrix, at first to carry out normalization and handle first test matrix, obtain second test matrix:

$A^{\′}={\left(\begin{array}{cccc}{x}_{11}^{\′}& x_{12}^{\′}& \·\·\·& x_{1m}^{\′}\\ x_{21}^{\′}& x_{22}^{\′}& \·\·\·& x_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ x_{n1}^{\′}& x_{n2}^{\′}& \·\·\·& x_{\mathrm{nm}}^{\′}\end{array}\right)}_{n\×m},$

Wherein, x ' _IjBe the x in first test matrix _IjNormalized value.Concrete normalized method can for: for the element x in first test matrix _Ij, confirm element x _IjGreatest member max and least member min in one row at place in all elements; Because x _IjBe the test value to i index, therefore consider the relation of i index and network equipment combination property, if the test value of this i index is big more, combination property is good more, then adopts formula

Confirm x _IjNormalized value x ' _IjIf the test value of this i index is more little, combination property is good more, then adopts formula Confirm x _IjNormalized value x ' _IjCertainly, can also adopt additive method that first test matrix is carried out normalization and handle, for example directly with x _IjWith the ratio of max as normalized value x ' _Ij

Behind second test matrix after obtaining normalization and handling, transposed matrix and this second test matrix of second test matrix multiplied each other, obtains the 3rd test matrix:

$B=A^{\′T}\×A^{\′}={\left(\begin{array}{cccc}{\mathrm{\α}}_{11}& {\mathrm{\α}}_{12}& \·\·\·& {\mathrm{\α}}_{1m}\\ {\mathrm{\α}}_{21}& {\mathrm{\α}}_{22}& \·\·\·& {\mathrm{\α}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\α}}_{m1}& {\mathrm{\α}}_{m2}& \·\·\·& {\mathrm{\α}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

According to the fiducial value that is each target setting, each fiducial value is added in said the 3rd test matrix accordingly, obtain the 4th test matrix:

$B={\left(\begin{array}{cccc}{\mathrm{\α}}_{11}& {\mathrm{\α}}_{12}& \·\·\·& {\mathrm{\α}}_{1m}\\ {\mathrm{\α}}_{21}& {\mathrm{\α}}_{22}& \·\·\·& {\mathrm{\α}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\α}}_{m1}& {\mathrm{\α}}_{m2}& \·\·\·& {\mathrm{\α}}_{\mathrm{mm}}\\ {\mathrm{\β}}_{m+\mathrm{1,1}}& {\mathrm{\β}}_{m+\mathrm{1,2}}& \·\·\·& {\mathrm{\β}}_{m+1,m}\end{array}\right)}_{(m+1)\×m},$

Wherein, β _{M+1, i}The fiducial value of each target setting is the fiducial value of i index in m the index, for can be set as required.

Said the 4th test matrix is carried out normalization handles, obtain the 5th test matrix:

$B^{\′}={\left(\begin{array}{cccc}{\mathrm{\α}}_{11}^{\′}& {\mathrm{\α}}_{12}^{\′}& \·\·\·& {\mathrm{\α}}_{1m}^{\′}\\ {\mathrm{\α}}_{21}^{\′}& {\mathrm{\α}}_{22}^{\′}& \·\·\·& {\mathrm{\α}}_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\α}}_{m1}^{\′}& {\mathrm{\α}}_{m2}^{\′}& \·\·\·& {\mathrm{\α}}_{\mathrm{mm}}^{\′}\\ {\mathrm{\β}}_{m+\mathrm{1,1}}^{\′}& {\mathrm{\β}}_{m+\mathrm{1,2}}^{\′}& \·\·\·& {\mathrm{\β}}_{m+1,m}^{\′}\end{array}\right)}_{(m+1)\×m},$

Wherein, α ' _IjBe the α in the 4th test matrix _IjNormalized value, β ' _{M+1, i}Be the β in the 4th test matrix _{M+1, i}Normalized value.To the 4th test matrix carry out method that normalization handles specifically can for: for the element α in the 4th test matrix _Ij, confirm element α _IjGreatest member max and least member min in one row at place in all elements, and adopt formula

Confirm α _IjNormalized value α ' _IjCertainly, also can adopt formula

or additive method to carry out normalization handles.

According to said the 5th test matrix, adopt following formula to confirm the distance of the fiducial value of each index and each index:

$d_{i,m+1}=\sqrt{\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{({\mathrm{\α}}_{\mathrm{ij}}^{\′}-{\mathrm{\β}}_{(m+1),j}^{\′})}^2},$

Wherein, i=1,2,3...m, d _{I, m+1}Distance for the fiducial value of i index in m the index and this i index;

To any two indexs, confirm the relative importance degree value between these two indexs based on following formula:

$r_{\mathrm{ij}}=\frac{d_{j,m+1}}{d_{i,m+1}+d_{j,m+1}},$

Wherein, r _IjBe the relative importance degree value of i index in m the index with respect to j index.This shows that for any two indexs of the network equipment, i index is with respect to the relative importance degree value r of j index _Ij, with the relative importance degree value r of j index with respect to i index _JiWith value be 1, and i index is 0.5 with respect to himself relative importance degree value of (i index).

In addition; Consider the difference of practical application scene; Relative importance degree value between each index of the network equipment can also be carried out artificial adjustment; So that last test result be to the test result of certain application-specific scene, the relative importance degree value of adjustment still need satisfy r _IjWith r _JiWith value be 1, and r _IiValue be 0.5.

Adopt said method to confirm after the relative importance degree value between each index; In step S103 shown in Figure 1; With the relative importance degree value between each index of confirming is that the method for element structure fuzzy evaluation matrix is specially, and constructs following matrix as the fuzzy evaluation matrix:

$R={\left(\begin{array}{cccc}{r}_{11}& r_{12}& ...& r_{1m}\\ r_{21}& r_{22}& ...& r_{2m}\\ ...& ...& ...& ...\\ r_{m1}& r_{m2}& ...& r_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, r _IjBe the relative importance degree value of i index with respect to j index, r _IjWith r _JiWith value be 1, and r _IiValue be 0.5, also promptly, the value of the diagonal entry among the fuzzy evaluation matrix R is 0.5, any two with respect to the elements of diagonal symmetry with value be 1.

Confirm that according to above-mentioned fuzzy evaluation matrix R the method for weight sets is specially: the fuzzy evaluation matrix R to structure carries out the fuzzy consensus processing, obtains Fuzzy consistent matrix M:

$M={\left(\begin{array}{cccc}{\mathrm{\ξ}}_{11}& {\mathrm{\ξ}}_{12}& ...& {\mathrm{\ξ}}_{1m}\\ {\mathrm{\ξ}}_{21}& {\mathrm{\ξ}}_{22}& ...& {\mathrm{\ξ}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{\ξ}}_{m1}& {\mathrm{\ξ}}_{m2}& ...& {\mathrm{\ξ}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, for the arbitrary element ξ among the Fuzzy consistent matrix M _Ij,

r _IkBe the element of the capable k row of i in the fuzzy evaluation matrix, r _JkElement for the capable k row of j in the fuzzy evaluation matrix;

According to Fuzzy consistent matrix M, adopt following formula to confirm the weight that each index is corresponding:

${\mathrm{\ω}}_i=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ξ}}_{\mathrm{ij}}-0.5}{\frac{m(m-1)}{2}},$

Wherein, ω _iBe the corresponding weight of i index, be the relative importance degree value of having taken all factors into consideration between each index after, this i index influences the degree value of network equipment combination property.

With the corresponding weight of confirming of each index is that element constitutes weight sets: ω=(ω ₁, ω ₂, ω ₃... ω _m).

In step S104 shown in Figure 1; The method of confirming the percentage of this index on each performance rate of test is specially: to each performance rate; In the corresponding a plurality of test values of this index of confirming to obtain; The number that belongs to the test value in the corresponding index test value scope of this performance rate, the ratio of the sum of a plurality of test values that the number of confirming is corresponding with this index that obtains is confirmed as the percentage of this index on this performance rate of test.

Wherein, this index can be confirmed according to before writing down in the test log this index is tested each test value that obtains at every turn to the corresponding index test value scope of each performance rate.For example; Before writing down in the test record, this index is tested the descending ordering of resulting each test value at every turn; The corresponding test value scope of test value that comes preceding 25% is confirmed as this index to the corresponding index test value scope of the first estate; The corresponding test value scope of test value that comes 25%～50% is confirmed as this index to the corresponding index test value scope of second grade, by that analogy, confirm the index test value scope that each performance rate is corresponding.

In addition; The percentage of this index on each performance rate can also adopt other modes to confirm; For example add up each user to this evaluation of indexes,, confirm with this index evaluation to be the number of this performance rate to each performance rate; With the ratio of the sum of the evaluation of number of confirming and statistics, confirm as the percentage of this index on this performance rate.

Is that the method that element structure is passed judgment on subclass is specially with this index of test at the percentage on each performance rate: according to the percentage of this index on each performance rate, construct following judge subclass:

Y _i＝(y _i1，y _i2，y _i3...y _ip)，

Wherein, Y _iBe the judge subclass of i index, p is the number of performance rate, for Y _iIn arbitrary element y _Iq, y _IqBe the percentage of this i index on q performance rate.

After adopting said method all to confirm corresponding judge subclass to each index, in step S105 shown in Figure 1, the judge matrix of confirming is specially: based on the judge subclass of confirming to each index respectively, confirm following judge matrix:

$Y={\left(\begin{array}{cccc}{y}_{11}& y_{12}& ...& y_{1p}\\ y_{21}& y_{22}& ...& y_{2p}\\ ...& ...& ...& ...\\ y_{m1}& y_{m2}& ...& y_{\mathrm{mp}}\end{array}\right)}_{m\×p},$

Wherein, for the arbitrary element y among the Y _Iq, y _IqBe the percentage of i index on q performance rate.

In step S106 shown in Figure 1, testing apparatus then can be with weight sets ω=(ω ₁, ω ₂, ω ₃... ω _m) and pass judgment on matrix $Y={\left(\begin{array}{cccc}{y}_{11}& y_{12}& ...& y_{1p}\\ y_{21}& y_{22}& ...& y_{2p}\\ ...& ...& ...& ...\\ y_{m1}& y_{m2}& ...& y_{\mathrm{mp}}\end{array}\right)}_{m\×p}$ Multiply each other, obtain performance and pass judgment on collection:

S＝ω×Y＝(s ₁，s ₂，s ₃...s _p)，

Wherein, performance is passed judgment on the percentage of combination property on each performance rate that each element that collects among the S is the network equipment, supposes the element s among the performance judge collection S _qBe the element of maximum, then with this element s _qQ corresponding performance rate confirmed as the test result that obtains.

Be example explanation test process with above-mentioned table 1 below.

The index of this network equipment comprises: index 1, index 2, index 3, index 4, testing apparatus carries out 100 tests respectively to these four indexs, obtains 100 corresponding respectively test values of these four indexs.Then construct first test matrix $A={\left(\begin{array}{cccc}{x}_{11}& x_{12}& x_{13}& x_{14}\\ x_{21}& x_{22}& x_{23}& x_{24}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ x_{\mathrm{100,1}}& x_{\mathrm{100,2}}& x_{\mathrm{100,3}}& x_{\mathrm{100,4}}\end{array}\right)}_{100\×4},$ First test matrix is carried out the normalization processing obtain second test matrix, and the transposed matrix and second test matrix of second test matrix multiplied each other, obtain the 3rd test matrix, suppose to be respectively β to the fiducial value of these four target settings ₅₁, β ₅₂, β ₅₃, β ₅₄, then add this four fiducial values, and carry out normalization and handle the 5th test matrix obtain and do $B^{\′}={\left(\begin{array}{cccc}{\mathrm{\α}}_{11}^{\′}& {\mathrm{\α}}_{12}^{\′}& {\mathrm{\α}}_{13}^{\′}& {\mathrm{\α}}_{14}^{\′}\\ {\mathrm{\α}}_{21}^{\′}& {\mathrm{\α}}_{22}^{\′}& {\mathrm{\α}}_{23}^{\′}& {\mathrm{\α}}_{24}^{\′}\\ {\mathrm{\α}}_{31}^{\′}& {\mathrm{\α}}_{32}^{\′}& {\mathrm{\α}}_{33}^{\′}& {\mathrm{\α}}_{34}^{\′}\\ {\mathrm{\α}}_{41}^{\′}& {\mathrm{\α}}_{42}^{\′}& {\mathrm{\α}}_{43}^{\′}& {\mathrm{\α}}_{44}^{\′}\\ {\mathrm{\β}}_{51}^{\′}& {\mathrm{\β}}_{52}^{\′}& {\mathrm{\β}}_{53}^{\′}& {\mathrm{\β}}_{54}^{\′}\end{array}\right)}_{5\×4},$ Suppose that the fuzzy evaluation matrix that obtains according to the 5th test matrix does $R={\left(\begin{array}{cccc}0.5& 0.3& 0.6& 0.8\\ 0.7& 0.5& 0.6& 0.9\\ 0.4& 0.4& 0.5& 0.6\\ 0.2& 0.1& 0.4& 0.5\end{array}\right)}_{4\×4},$ The Fuzzy consistent matrix that then processing obtains through fuzzy consensus does $M={\left(\begin{array}{cccc}0.5& 0.4375& 0.5375& 0.625\\ 0.5625& 0.5& 0.6& 0.6875\\ 0.4625& 0.4& 0.5& 0.5875\\ 0.375& 0.3125& 0.4125& 0.5\end{array}\right)}_{4\×4},$ And then the weight sets that obtains is ω=(0.2667; 0.3083; 0.2417; 0.1833), suppose that the performance rate of setting is respectively: the first estate, second grade, the tertiary gradient, the fourth estate, the judge matrix of confirming according to the percentage of confirming to these four indexs on each performance rate does $Y={\left(\begin{array}{cccc}0.1& 0.1& 0.8& 0\\ 0& 0.2& 0.7& 0.1\\ 0.1& 0.2& 0.5& 0.2\\ 0.2& 0.1& 0.6& 0.1\end{array}\right)}_{4\×4},$ Wherein, the value 0.1 of passing judgment on the 1st row the 1st row in the matrix is the percentage of index 1 on the first estate, and the value 0.1 of the 1st row the 2nd row is the percentage of index 1 on second grade; The value 0 of the 2nd row the 1st row is the percentage of index 2 on the first estate, and by that analogy, the performance that then weight sets and judge matrix multiple is obtained is passed judgment on collection and is S=ω * Y=(0.0875; 0.155,0.66,0.0975); It is thus clear that it is greatest member that performance is passed judgment on the element of concentrating 0.66; Its corresponding performance rate is the tertiary gradient, therefore with this tertiary gradient as the network equipment being tested resulting test result, this test result is the combination property of this network equipment of test.

In embodiments of the present invention, consider each index that also possibly be directed against the network equipment in the practical application, this index is subdivided into more fine-grained subordinate index again.For example, the index of router comprises throughput, time delay etc., and throughput and time delay are the first class index of router.Throughput is divided into again: device throughput and port throughput, device throughput and port throughput affect the throughput of router jointly, so these two indexs are the two-level index under the throughput.Time delay is divided into again: maximum delay, minimal time delay, average delay, these three indexs affect the time delay of router jointly, are the two-level index under the time delay.Therefore, based on the thinking same with method of testing shown in Figure 1, the present invention also provides another kind of method of testing, is used for the combination property of the network equipment that is subdivided into the firsts and seconds index is tested.

Fig. 2 is the another kind of test process that the embodiment of the invention provides, and specifically may further comprise the steps:

S201: each two-level index to the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each two-level index.

In embodiments of the present invention, the number of times that testing apparatus is tested each two-level index also can be set as required, for example 100 times.

Wherein, the hierarchical relationship of the first class index of this network equipment and two-level index is as shown in table 2.

Table 2

As shown in table 2, the first class index of this network equipment comprises: index 1, index 2, index 3, index 4 comprise several two-level index again under each first class index.With index 1 is example, and each two-level index that index comprises for 1 time is respectively: index 1 ₁, index 1 ₂, index 1 ₃, index 1 ₄, these four two-level index affect index 1 jointly, and the relation of other first class index and two-level index is also basic identical with above-mentioned relation.Therefore, testing apparatus is repeatedly tested respectively each two-level index, for example respectively each two-level index is carried out 100 tests, obtains 100 corresponding respectively test values of each two-level index.

S202:,, confirm the relative importance degree value between each two-level index under this first class index according to the corresponding a plurality of test values of each two-level index under this first class index that obtains to each first class index of this network equipment.

In embodiments of the present invention, testing apparatus can be directed against each first class index, according to the corresponding a plurality of test values of each two-level index under this first class index that obtains, confirms the relative importance degree value between each two-level index under this first class index.Relative importance degree value between each two-level index under this first class index is represented: each two-level index under this first class index is compared in twos, and the degree that one of them two-level index influences this first class index influences the comparison value of the degree of this first class index than another and index.

S203: with the relative importance degree value between each two-level index of confirming is the corresponding secondary fuzzy evaluation matrix of this first class index of element structure, according to the secondary fuzzy evaluation matrix of structure, confirms the secondary weight sets that this first class index is corresponding.

After the relative importance degree value between each two-level index under this first class index that testing apparatus is confirmed; Be element structure fuzzy evaluation matrix then with each the relative importance degree value between each two-level index; And the corresponding secondary weight sets of definite in view of the above this first class index; Each element representation in the secondary weight sets of this first class index correspondence: after having considered the relative importance degree value between each two-level index under this first class index, each two-level index under this first class index influences the degree value of this first class index.

S204: to each two-level index under this first class index; According to the corresponding a plurality of test values of this two-level index that obtain; And the corresponding two-level index test value scope of setting of each performance rate; Confirming the percentage of this two-level index on each performance rate of test, is that the corresponding secondary of this two-level index of element structure is passed judgment on subclass with the percentage of this two-level index on each performance rate of testing.

In embodiments of the present invention, can limit each performance rate in advance, for example be set at four performance rates, be respectively: the first estate, second grade, the tertiary gradient, the fourth estate.And can set the corresponding index test value scope of each performance rate to each two-level index under this first class index.Testing apparatus then can be to each two-level index under this first class index; Confirm in the corresponding test value of this two-level index; Drop on the quantity of the test value in the corresponding index test value scope of each performance rate respectively; And then confirm the percentage of this two-level index on each performance rate, and construct the corresponding secondary of this two-level index in view of the above and pass judgment on subclass.

S205: according to passing judgment on subclass to the secondary of each the two-level index structure under this first class index respectively; Confirm the secondary judge matrix that this first class index is corresponding; The secondary that the secondary weight sets that this first class index is corresponding is corresponding with this first class index is passed judgment on matrix multiple, obtains the corresponding one-level of this first class index and passes judgment on collection.

In embodiments of the present invention; After having confirmed the secondary judge subclass of each two-level index correspondence under this first class index respectively; The element of then passing judgment in the subclass with each secondary is confirmed the secondary judge matrix that this first class index is corresponding, and this secondary is passed judgment on the corresponding secondary of each two-level index that has comprised in matrix under this first class index and passed judgment on the element in the subclass.

Because each two-level index under this first class index of each element representation in the secondary weight sets of this first class index correspondence influences the degree value of this first class index; The secondary that this first class index is corresponding is passed judgment on the percentage of certain two-level index on certain performance rate under this first class index of each element representation in the matrix; Therefore the secondary that this first class index is corresponding secondary weight sets is corresponding with this first class index is passed judgment on matrix multiple, and the one-level that resulting this first class index is corresponding is passed judgment on each element of concentrating and is: the percentage of this first class index on each performance rate.

S206: the one-level based on confirming to each first class index of this network equipment is respectively passed judgment on collection, confirms one-level judge matrix.

In embodiments of the present invention; After testing apparatus has confirmed all that to each first class index the corresponding one-level of this first class index is passed judgment on collection respectively; Then pass judgment on the element of concentrating with each one-level and confirm one-level judge matrix, this one-level is passed judgment on and has been comprised all elements that each one-level judge is concentrated in matrix.

S207: confirm the relative importance degree value between each first class index of this network equipment; With the relative importance degree value between each first class index is element structure one-level fuzzy evaluation matrix, confirms the one-level weight sets according to the one-level fuzzy evaluation matrix of structure.

In embodiments of the present invention; Testing apparatus also will be confirmed the relative importance degree value between each first class index of the network equipment; Relative importance degree value between each first class index is represented: each first class index of this network equipment is compared in twos, and the degree that one of them first class index influences network equipment combination property influences the comparison value of the degree of network equipment combination property than another index.

After having confirmed the relative importance degree value between each first class index; Be element structure one-level fuzzy evaluation matrix then with each relative importance degree value; And definite in view of the above one-level weight sets; Each element representation in this one-level weight sets: after having taken all factors into consideration the relative importance degree value between each first class index, each first class index influences the degree value of network equipment combination property accordingly.

S208: one-level weight sets of confirming and one-level are passed judgment on matrix multiple, and the performance that obtains this network equipment is passed judgment on collection, this performance is passed judgment on concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

Because each first class index of each element representation in the one-level weight sets influences the degree value of network equipment combination property; One-level is passed judgment on the percentage of certain first class index of each element representation on certain performance rate in the matrix; Therefore one-level weight sets and one-level are passed judgment on matrix multiple; Resulting performance is passed judgment on the percentage of combination property on each performance rate that each element of concentrating is the network equipment, therefore performance is passed judgment on and is concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

In said process; Testing apparatus is repeatedly tested each two-level index of the network equipment; Obtain the corresponding a plurality of test values of each two-level index, to each first class index, according to the corresponding secondary fuzzy evaluation matrix of this first class index of the structure of the relative importance degree value between each two-level index under this first class index; Confirm the secondary weight sets that this first class index is corresponding according to the secondary fuzzy evaluation matrix of structure; And the percentage of each two-level index on each performance rate under this first class index of confirming to test, confirm the secondary judge matrix that this first class index is corresponding in view of the above, the secondary judge matrix that the secondary weight sets that this first class index is corresponding is corresponding with this first class index equates; Obtain the corresponding one-level of this first class index and pass judgment on collection; Pass judgment on collection according to the one-level that obtains to each first class index respectively and confirm one-level judge matrix,, confirm the one-level weight sets in view of the above according to the structure of the relative importance degree value between each first class index one-level fuzzy evaluation matrix; One-level weight sets and one-level are passed judgment on the performance judge collection that matrix multiple obtains this network equipment, and the performance rate that wherein maximum element is corresponding is confirmed as the test result that obtains.Pass through said method; Testing apparatus is all tested each two-level index; And combine of the influence of each two-level index to the first class index under it; And each first class index is upwards confirmed the combination property of the network equipment step by step, so can be tested accurately the combination property of the network equipment the influence of the combination property of the network equipment.

When the index of the network equipment is subdivided into three grades, level Four or when more multistage, method shown in Figure 2 stands good.With three grades is example; Can adopt similar method with Fig. 2; Each three grades of index is repeatedly tested; And comprehensive each three grades of index to the influence of the two-level index under it, each two-level index to the influence of the first class index under it, each first class index influence to the combination property of the network equipment, upwards confirm the combination property of the network equipment step by step, just give unnecessary details no longer one by one here.

In step S201 shown in Figure 2, testing apparatus carries out the test of set point number according to set point number to each two-level index of the network equipment, obtains the corresponding a plurality of test values of each two-level index.Wherein, to each two-level index, the quantity of a plurality of test values that this two-level index that obtains is corresponding equates with this set point number.And in order further to improve the accuracy of test, it is bigger that this set point number can be set, preferable, and this set point number is not less than the number of the two-level index of the network equipment.

In step S202 shown in Figure 2; Testing apparatus is to each first class index of the network equipment; According to the corresponding a plurality of test values of each two-level index under this first class index that obtains; The method of confirming the relative significance level value between each two-level index under this first class index is specially: to each first class index of the network equipment, be that element is constructed first test matrix with the corresponding a plurality of test values of each two-level index under this first class index that obtains:

$\mathrm{Ag}={\left(\begin{array}{cccc}{\mathrm{xg}}_{11}& {\mathrm{xg}}_{12}& \·\·\·& {\mathrm{xg}}_{1m}\\ {\mathrm{xg}}_{21}& {\mathrm{xg}}_{22}& \·\·\·& {\mathrm{xg}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{xg}}_{n1}& {\mathrm{xg}}_{n2}& \·\·\·& {\mathrm{xg}}_{\mathrm{nm}}\end{array}\right)}_{n\×m},$

Wherein, n is said set point number, and m is the number of the two-level index under this first class index, xg _IjBe the i time resulting test value of test that j two-level index under g the first class index carried out; First test matrix according to structure is confirmed the relative importance degree value between each two-level index under this first class index.

When confirming the relative importance degree value between each two-level index under this first class index according to above-mentioned first test matrix, at first to carry out normalization and handle first test matrix, obtain second test matrix:

${\mathrm{Ag}}^{\′}={\left(\begin{array}{cccc}{\mathrm{xg}}_{11}^{\′}& {\mathrm{xg}}_{12}^{\′}& \·\·\·& {\mathrm{xg}}_{1m}^{\′}\\ {\mathrm{xg}}_{21}^{\′}& {\mathrm{xg}}_{22}^{\′}& \·\·\·& {\mathrm{xg}}_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{xg}}_{n1}^{\′}& {\mathrm{xg}}_{n2}^{\′}& \·\·\·& {\mathrm{xg}}_{\mathrm{nm}}^{\′}\end{array}\right)}_{n\×m},$

Wherein, xg ' _IjBe the xg in first test matrix _IjNormalized value.Concrete normalized method can for: for the element x g in first test matrix _Ij, confirm element x g _IjGreatest member max and least member min in one row at place in all elements; Because xg _IjBe the i time resulting test value of test that j two-level index under g the first class index carried out; Therefore consider the relation of this j two-level index and g first class index; If the test value of this j two-level index is big more, g first class index is good more, then adopts formula

Confirm xg _IjNormalized value xg ' _IjIf the test value of this j index is more little, g first class index is good more, then adopts formula

Confirm xg _IjNormalized value xg ' _IjCertainly, can also adopt additive method that first test matrix is carried out normalization and handle, for example directly with xg _IjWith the ratio of max as normalized value xg ' _Ij

Behind second test matrix after obtaining normalization and handling, the transposed matrix and second test matrix of second test matrix multiplied each other, obtains the 3rd test matrix:

$\mathrm{Bg}={\mathrm{Ag}}^{\′T}\×{\mathrm{Ag}}^{\′}={\left(\begin{array}{cccc}{\mathrm{\αg}}_{11}& {\mathrm{\αg}}_{12}& \·\·\·& {\mathrm{\αg}}_{1m}\\ {\mathrm{\αg}}_{21}& {\mathrm{\αg}}_{22}& \·\·\·& {\mathrm{\αg}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\αg}}_{m1}& {\mathrm{\αg}}_{m2}& \·\·\·& {\mathrm{\αg}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

According to the fiducial value of setting for each two-level index under this first class index, each fiducial value is added in said the 3rd test matrix accordingly, obtain the 4th test matrix:

$\mathrm{Bg}={\left(\begin{array}{cccc}{\mathrm{\αg}}_{11}& {\mathrm{\αg}}_{12}& \·\·\·& {\mathrm{\αg}}_{1m}\\ {\mathrm{\αg}}_{21}& {\mathrm{\αg}}_{22}& \·\·\·& {\mathrm{\αg}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\αg}}_{m1}& {\mathrm{\αg}}_{m2}& \·\·\·& {\mathrm{\αg}}_{\mathrm{mm}}\\ {\mathrm{\βg}}_{m+\mathrm{1,1}}& {\mathrm{\βg}}_{m+\mathrm{1,2}}& \·\·\·& {\mathrm{\βg}}_{m+1,m}\end{array}\right)}_{(m+1)\×m},$

Wherein, β g _{M+1, i}The fiducial value that each two-level index is set is the fiducial value of i two-level index under this g first class index, for can be set as required.

Said the 4th test matrix is carried out normalization handles, obtain the 5th test matrix:

${\mathrm{Bg}}^{\′}={\left(\begin{array}{cccc}{\mathrm{\αg}}_{11}^{\′}& {\mathrm{\αg}}_{12}^{\′}& \·\·\·& {\mathrm{\αg}}_{1m}^{\′}\\ {\mathrm{\αg}}_{21}^{\′}& {\mathrm{\αg}}_{22}^{\′}& \·\·\·& {\mathrm{\αg}}_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\αg}}_{m1}^{\′}& {\mathrm{\αg}}_{m2}^{\′}& \·\·\·& {\mathrm{\αg}}_{\mathrm{mm}}^{\′}\\ {\mathrm{\βg}}_{m+\mathrm{1,1}}^{\′}& {\mathrm{\βg}}_{m+\mathrm{1,2}}^{\′}& \·\·\·& {\mathrm{\βg}}_{m+1,m}^{\′}\end{array}\right)}_{(m+1)\×m},$

Wherein, α g ' _IjBe the α g in the 4th test matrix _IjNormalized value, β g ' _{M+1, i}Be the β g in the 4th test matrix _{M+1, i}Normalized value.To the 4th test matrix carry out method that normalization handles specifically can for: for the element α g in the 4th test matrix _Ij, confirm element α g _IjGreatest member max and least member min in one row at place in all elements, and adopt formula

Confirm α g _IjNormalized value α g ' _IjCertainly, also can adopt formula

or additive method to carry out normalization handles.

According to said the 5th test matrix, adopt following formula to confirm the distance of the fiducial value of each two-level index and each two-level index under this first class index:

${\mathrm{dg}}_{i,m+1}=\sqrt{\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{({\mathrm{\αg}}_{\mathrm{ij}}^{\′}-{\mathrm{\βg}}_{(m+1),j}^{\′})}^2},$

Wherein, i=1,2,3...m, dg _{I, m+1}Distance for the fiducial value of i two-level index under this g first class index and this i two-level index;

To any two two-level index under this first class index, confirm the relative importance degree value between these two two-level index based on following formula:

${\mathrm{rg}}_{\mathrm{ij}}=\frac{{\mathrm{dg}}_{j,m+1}}{{\mathrm{dg}}_{i,m+1}+{\mathrm{dg}}_{j,m+1}},$

Wherein, rg _IjBe the relative importance degree value of i two-level index under this g first class index with respect to j two-level index under this g first class index.This shows that for any two two-level index under this g first class index, i two-level index is with respect to the relative importance degree value rg of j two-level index _Ij, with the relative importance degree value rg of j two-level index with respect to i two-level index _JiWith value be 1, and i two-level index is 0.5 with respect to himself relative importance degree value of (i two-level index).

In addition; Consider the difference of practical application scene; Relative importance degree value between each two-level index under this first class index can also be carried out artificial adjustment; So that last test result be to the test result of certain application-specific scene, the relative importance degree value of adjustment still need satisfy rg _IjWith rg _JiWith value be 1, and rg _IiValue be 0.5.

Adopt said method to confirm after the relative importance degree value between each two-level index under this first class index; In step S203 shown in Figure 2; With the relative importance degree value between each two-level index of confirming is that the method for the corresponding secondary fuzzy evaluation matrix of this first class index of element structure is specially, and constructs following matrix and is the corresponding secondary fuzzy evaluation matrix of this first class index:

$\mathrm{Rg}={\left(\begin{array}{cccc}{\mathrm{rg}}_{11}& {\mathrm{rg}}_{12}& ...& {\mathrm{rg}}_{1m}\\ {\mathrm{rg}}_{21}& {\mathrm{rg}}_{22}& ...& {\mathrm{rg}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{rg}}_{m1}& {\mathrm{rg}}_{m2}& ...& {\mathrm{rg}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, rg _IjBe the relative importance degree value of i two-level index under this g first class index with respect to j two-level index under this g first class index, rg _IjWith rg _JiWith value be 1, and rg _IiValue be 0.5, also promptly, the value of the diagonal entry among the corresponding secondary fuzzy evaluation matrix Rg of this g first class index is 0.5, any two with respect to the elements of diagonal symmetry with value be 1.

According to above-mentioned secondary fuzzy evaluation matrix Rg, confirm that the method for the secondary weight sets that this first class index is corresponding is specially: the secondary fuzzy evaluation matrix to structure carries out the fuzzy consensus processing, obtains secondary Fuzzy consistent matrix Mg:

$\mathrm{Mg}={\left(\begin{array}{cccc}{\mathrm{\ξg}}_{11}& {\mathrm{\ξg}}_{12}& ...& {\mathrm{\ξg}}_{1m}\\ {\mathrm{\ξg}}_{21}& {\mathrm{\ξg}}_{22}& ...& {\mathrm{\ξg}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{\ξg}}_{m1}& {\mathrm{\ξg}}_{m2}& ...& {\mathrm{\ξg}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, for the arbitrary element ξ g among the secondary Fuzzy consistent matrix Mg _Ij, ${\mathrm{\ξg}}_{\mathrm{ij}}=\frac{\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{rg}}_{\mathrm{ik}}-\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{rg}}_{\mathrm{jk}}}{2m}+0.5,$ Rg _IkBe the element of the capable k row of i in the secondary fuzzy evaluation matrix, rg _JkElement for the capable k row of j in the secondary fuzzy evaluation matrix;

According to secondary Fuzzy consistent matrix Mg, adopt following formula to confirm the corresponding weight of each two-level index under this first class index:

${\mathrm{\ωg}}_i=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ξg}}_{\mathrm{ij}}-0.5}{\frac{m(m-1)}{2}},$

Wherein, ω g _iBe the corresponding weight of i two-level index under this g first class index;

With the corresponding weight of each two-level index under this first class index of confirming is that element constitutes the corresponding secondary weight sets of this first class index: ω g=(ω g ₁, ω g ₂, ω g ₃... ω g _m).

In step S204 shown in Figure 2; The method of confirming the percentage of this two-level index on each performance rate of test is specially: to each performance rate; In the corresponding a plurality of test values of this two-level index of confirming to obtain; The number that belongs to the test value in the corresponding index test value scope of this performance rate, the ratio of the sum of a plurality of test values that the number of confirming is corresponding with this two-level index that obtains is confirmed as the percentage of this two-level index on this performance rate of test.Wherein, this two-level index can be confirmed based on before writing down in the test log this two-level index is tested each test value that obtains at every turn to the corresponding index test value scope of each performance rate, just give unnecessary details no longer one by one here.

In addition; The percentage of this two-level index on each performance rate can also adopt other modes to confirm; For example add up the evaluation of each user,, confirm this two-level index is evaluated as the number of this performance rate to each performance rate to this two-level index; With the ratio of the sum of the evaluation of number of confirming and statistics, confirm as the percentage of this two-level index on this performance rate.

Is that the method that the corresponding secondary of this two-level index of element structure is passed judgment on subclass is specially with this two-level index of test at the percentage on each performance rate: according to the percentage of this two-level index on each performance rate, construct following secondary and pass judgment on subclass:

Yg _i＝(yg _i1，yg _i2，yg _i3...yg _ip)，

Wherein, Yg _iThe secondary that is i two-level index under g the first class index is passed judgment on subclass, and p is the number of performance rate, for Yg _iIn arbitrary element yg _Iq, yg _IqBe the percentage of i two-level index on q performance rate under this g first class index.

After adopting said method to confirm all that to each two-level index under this first class index corresponding secondary is passed judgment on subclass; In step S205 shown in Figure 2; The secondary of this first class index correspondence of confirming is passed judgment on matrix and is specially: the secondary according to confirming to each two-level index under this first class index is respectively passed judgment on subclass, confirms following secondary judge matrix:

$\mathrm{Yg}={\left(\begin{array}{cccc}{\mathrm{yg}}_{11}& {\mathrm{yg}}_{12}& ...& {\mathrm{yg}}_{1p}\\ {\mathrm{yg}}_{21}& {\mathrm{yg}}_{22}& ...& {\mathrm{yg}}_{2p}\\ ...& ...& ...& ...\\ {\mathrm{yg}}_{m1}& {\mathrm{yg}}_{m2}& ...& {\mathrm{yg}}_{\mathrm{mp}}\end{array}\right)}_{m\×p},$

Wherein, for the arbitrary element yg among the Yg _Iq, yg _IqBe the percentage of i two-level index on q performance rate under this g first class index.

After having confirmed that the corresponding secondary of above-mentioned this first class index is passed judgment on matrix, secondary weight sets ω g=(the ω g that testing apparatus is then corresponding with this first class index ₁, ω g ₂, ω g ₃... ω g _m) the secondary judge matrix corresponding with this first class index $\mathrm{Yg}={\left(\begin{array}{cccc}{\mathrm{Yg}}_{11}& {\mathrm{Yg}}_{12}& ...& {\mathrm{Yg}}_{1p}\\ {\mathrm{Yg}}_{21}& {\mathrm{Yg}}_{22}& ...& {\mathrm{Yg}}_{2p}\\ ...& ...& ...& ...\\ {\mathrm{Yg}}_{m1}& {\mathrm{Yg}}_{m2}& ...& {\mathrm{Yg}}_{\mathrm{Mp}}\end{array}\right)}_{m\×p},$ Multiply each other, the one-level judge collection that obtains this first class index correspondence is: $\mathrm{Sg}=({\mathrm{\ω\; g}}_1,{\mathrm{\ω\; g}}_2,{\mathrm{\ω\; g}}_3...{\mathrm{\ω\; g}}_m)\×{\left(\begin{array}{cccc}{\mathrm{Yg}}_{11}& {\mathrm{Yg}}_{12}& ...& {\mathrm{Yg}}_{1p}\\ {\mathrm{Yg}}_{21}& {\mathrm{Yg}}_{22}& ...& {\mathrm{Yg}}_{2p}\\ ...& ...& ...& ...\\ {\mathrm{Yg}}_{m1}& {\mathrm{Yg}}_{m2}& ...& {\mathrm{Yg}}_{\mathrm{Mp}}\end{array}\right)}_{m\×p}=({\mathrm{Sg}}_1,{\mathrm{Sg}}_2,{\mathrm{Sg}}_3...{\mathrm{Sg}}_p),$

Wherein, Sg is that the corresponding one-level of g first class index is passed judgment on collection, ω g=(ω g ₁, ω g ₂, ω g ₃... ω g _m) be the corresponding secondary weight sets of confirming of this g first class index, sg ₁, sg ₂, sg ₃... sg _pJudge weight for this g the corresponding p of first class index difference performance rate.

After testing apparatus has confirmed all that to each first class index corresponding one-level is passed judgment on collection, confirm that following matrix passes judgment on matrix as one-level:

$S={\left(\begin{array}{cccccc}{s1}_1& {s1}_2& ...& {s1}_q& ...& {s1}_p\\ {s2}_1& {s2}_2& ...& {s2}_q& ...& {s2}_p\\ ...& ...& ...& ...& ...& ...\\ {\mathrm{sg}}_1& {\mathrm{sg}}_2& ...& {\mathrm{sg}}_q& ...& {\mathrm{sg}}_p\\ ...& ...& ...& ...& ...& ...\\ {\mathrm{sh}}_1& {\mathrm{sh}}_2& ...& {\mathrm{sh}}_q& ...& {\mathrm{sh}}_p\end{array}\right)}_{h\×p},$

Wherein, S passes judgment on matrix for the one-level of confirming, h is the number of the first class index of the said network equipment, passes judgment on the arbitrary element sg in the matrix S for one-level _q, sg _qJudge weight for corresponding q the performance rate of g first class index in h the first class index of the said network equipment.

In step S207 shown in Figure 2, testing apparatus also will be confirmed the relative importance degree value between each first class index of the network equipment, and constructs one-level fuzzy evaluation matrix in view of the above:

$R={\left(\begin{array}{cccc}{r}_{11}& r_{12}& ...& r_{1h}\\ r_{21}& r_{22}& ...& r_{2h}\\ ...& ...& ...& ...\\ r_{h1}& r_{h2}& ...& r_{\mathrm{hh}}\end{array}\right)}_{h\×h},$

Wherein, h is the number of the first class index of the network equipment, r _IjBe the relative importance degree value of i first class index in h the first class index of the network equipment with respect to j first class index, r _IjWith r _JiWith value be 1, and r _IiValue be 0.5.

And; Any two first class index to the network equipment; I first class index can be the operation relation of confirming each two-level index under i first class index and this i first class index with respect to definite method of the relative importance degree value of j first class index; The test value of each two-level index under this i first class index of test is at every turn carried out computing according to the operation relation of confirming, obtain the test value of the corresponding each test of i first class index; After adopting said method to confirm the test value of the corresponding each test of each first class index; Can adopt and confirm the similar method of relative importance degree value between each two-level index under certain first class index; Test value according to the corresponding each test of each first class index; Confirm the relative importance degree value between each first class index, and based on the structure of the relative importance degree value between each first class index one-level fuzzy evaluation matrix.

For example; Each two-level index under the first class index A1 is: a1 and a2, and the operation relation of this first class index and each two-level index is: A1=a1+a2, then will test the test value that obtains to a1 for the first time; With test value that for the first time test obtains to a2 and value; As the corresponding test value of test for the first time of the first class index A1 that obtains, by that analogy, also be about to the test value that test obtains to a1 at every turn; With test value that at every turn test obtains to a2 and value, as the corresponding test value of testing of the first class index A1 that obtains at every turn.After having confirmed the test value of the corresponding each test of each first class index,, confirm the relative importance degree value between each first class index according to the test value of the corresponding each test of each first class index, and structure one-level fuzzy evaluation matrix.

Can also artificially set the relative importance degree value between each first class index as required, and based on the relative importance degree value between each first class index of artificial setting, structure one-level fuzzy evaluation matrix.

Wherein, No matter adopt which kind of method to confirm the relative importance degree value between each first class index; The relative importance degree value of confirming all need satisfy: i first class index is with respect to the relative importance degree value of j first class index; With j first class index with respect to the relative importance degree value of i first class index with value be 1, and i first class index is 0.5 with respect to himself relative importance degree value of (i first class index).

Testing apparatus confirms that the method for one-level weight sets is specially after having confirmed one-level fuzzy evaluation matrix R:

One-level fuzzy evaluation matrix to structure carries out the fuzzy consensus processing, obtains one-level Fuzzy consistent matrix M:

$M={\left(\begin{array}{cccc}{\mathrm{\ξ}}_{11}& {\mathrm{\ξ}}_{12}& ...& {\mathrm{\ξ}}_{1h}\\ {\mathrm{\ξ}}_{21}& {\mathrm{\ξ}}_{22}& ...& {\mathrm{\ξ}}_{2h}\\ ...& ...& ...& ...\\ {\mathrm{\ξ}}_{h1}& {\mathrm{\ξ}}_{h2}& ...& {\mathrm{\ξ}}_{\mathrm{hh}}\end{array}\right)}_{h\×h},$

Wherein, for the arbitrary element ξ among the one-level Fuzzy consistent matrix M _Ij,

r _IkBe the element of the capable k row of i in the one-level fuzzy evaluation matrix, r _JkElement for the capable k row of j in the one-level fuzzy evaluation matrix;

According to one-level Fuzzy consistent matrix M, adopt following formula to confirm the weight that each first class index is corresponding:

${\mathrm{\ω}}_i=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ξ}}_{\mathrm{ij}}-0.5}{\frac{h(h-1)}{2}},$

Wherein, ω _iBe the corresponding weight of this i first class index;

With the corresponding weight of confirming of each first class index is that element constitutes the one-level weight sets: ω=(ω ₁, ω ₂, ω ₃... ω _h).

It is thus clear that, confirm the method for one-level weight sets according to one-level fuzzy evaluation matrix, and confirm that according to the corresponding secondary fuzzy evaluation matrix of certain first class index the method for the secondary weight sets that this first class index is corresponding is basic identical.

After testing apparatus had confirmed that one-level is passed judgment on matrix S and one-level weight sets ω, in step S208 shown in Figure 2, testing apparatus then can be with one-level weight sets ω=(ω ₁, ω ₂, ω ₃... ω _h) pass judgment on matrix with one-level $S={\left(\begin{array}{cccccc}{s1}_1& {s1}_2& ...& {s1}_q& ...& {s1}_p\\ {s2}_1& {s2}_2& ...& {s2}_q& ...& {s2}_p\\ ...& ...& ...& ...& ...& ...\\ {\mathrm{sg}}_1& {\mathrm{sg}}_2& ...& {\mathrm{sg}}_q& ...& {\mathrm{sg}}_p\\ ...& ...& ...& ...& ...& ...\\ {\mathrm{sh}}_1& {\mathrm{sh}}_2& ...& {\mathrm{sh}}_q& ...& {\mathrm{sh}}_p\end{array}\right)}_{h\×p},$ Multiply each other, obtain performance and pass judgment on collection: T=ω * S=(t ₁, t ₂, t ₃... t _p), wherein, performance is passed judgment on the percentage of combination property on each performance rate that each element that collects among the T is the network equipment, supposes the element t among the performance judge collection T _qBe greatest member, then with this element t _qQ corresponding performance rate confirmed as the test result that obtains.

Be example explanation test process with table 2 below.

The first class index of this network equipment comprises: index 1, index 2, index 3, index 4, comprise several two-level index under each first class index, and testing apparatus carries out 100 tests respectively to each two-level index.Wherein, the two-level index under the index 1 is: index 1 ₁, index 1 ₂, index 1 ₃, index 1 ₄, then testing apparatus is that first test matrix that element is constructed does with the corresponding a plurality of test values of each two-level index under this index 1 that obtains $A1={\left(\begin{array}{cccc}{x1}_{11}& {x1}_{12}& {x1}_{13}& {x1}_{14}\\ {x1}_{21}& {x1}_{22}& {x1}_{23}& {x1}_{24}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {x1}_{\mathrm{100,1}}& {x1}_{\mathrm{100,2}}& {x1}_{\mathrm{100,3}}& {x1}_{\mathrm{100,4}}\end{array}\right)}_{100\×4}.$

First test matrix is carried out the normalization processing obtain second test matrix, and the transposed matrix and second test matrix of second test matrix multiplied each other, obtain the 3rd test matrix, suppose to index 1 ₁, index 1 ₂, index 1 ₃, index 1 ₄The fiducial value of setting respectively is β 1 ₅₁, β 1 ₅₂, β 1 ₅₃, β 1 ₅₄, then add this four fiducial values, and carry out normalization and handle the 5th test matrix obtain and do ${B1}^{\′}={\left(\begin{array}{cccc}{\mathrm{\α}1}_{11}^{\′}& {\mathrm{\α}1}_{12}^{\′}& {\mathrm{\α}1}_{13}^{\′}& {\mathrm{\α}1}_{14}^{\′}\\ {\mathrm{\α}1}_{21}^{\′}& {\mathrm{\α}1}_{22}^{\′}& {\mathrm{\α}1}_{23}^{\′}& {\mathrm{\α}1}_{24}^{\′}\\ {\mathrm{\α}1}_{31}^{\′}& {\mathrm{\α}1}_{32}^{\′}& {\mathrm{\α}1}_{33}^{\′}& {\mathrm{\α}1}_{34}^{\′}\\ {\mathrm{\α}1}_{41}^{\′}& {\mathrm{\α}1}_{42}^{\′}& {\mathrm{\α}1}_{43}^{\′}& {\mathrm{\α}1}_{44}^{\′}\\ {\mathrm{\β}1}_{51}^{\′}& {\mathrm{\β}1}_{52}^{\′}& {\mathrm{\β}1}_{53}^{\′}& {\mathrm{\β}1}_{54}^{\′}\end{array}\right)}_{5\×4}.$

Suppose that the index 1 corresponding secondary fuzzy evaluation matrix that obtains according to the 5th test matrix does $R1={\left(\begin{array}{cccc}0.5& 0.3& 0.6& 0.8\\ 0.7& 0.5& 0.6& 0.9\\ 0.4& 0.4& 0.5& 0.6\\ 0.2& 0.1& 0.4& 0.5\end{array}\right)}_{4\×4},$ This index 1 corresponding secondary Fuzzy consistent matrix that then processing obtains through fuzzy consensus does $M1={\left(\begin{array}{cccc}0.5& 0.4375& 0.5375& 0.625\\ 0.5625& 0.5& 0.6& 0.6875\\ 0.4625& 0.4& 0.5& 0.5875\\ 0.375& 0.3125& 0.4125& 0.5\end{array}\right)}_{4\×4},$ And then this index 1 corresponding secondary weight sets that obtains is ω 1=(0.2667,0.3083,0.2417,0.1833).

Suppose that the performance rate of setting is respectively: the first estate, second grade, the tertiary gradient, the fourth estate, the secondary judge matrix of confirming according to the percentage of confirming to four two-level index under the index 1 on each performance rate does $Y1={\left(\begin{array}{cccc}0.1& 0.1& 0.8& 0\\ 0& 0.2& 0.7& 0.1\\ 0.1& 0.2& 0.5& 0.2\\ 0.2& 0.1& 0.6& 0.1\end{array}\right)}_{4\×4},$ Wherein, the value 0.1 of the 1st row the 1st row in the secondary judge matrix is an index 1 ₁Percentage on the first estate, the value 0.1 of the 1st row the 2nd row is an index 1 ₁Percentage on second grade, the value 0 of the 2nd row the 1st row is an index 1 ₂Percentage on the first estate by that analogy, then obtains index 1 corresponding one-level with index 1 corresponding secondary weight sets with index 1 corresponding secondary judge matrix multiple and passes judgment on collection S1=ω 1 * Y1=(0.0875,0.155,0.66,0.0975).

Testing apparatus all carries out above-mentioned processing to index 2, index 3, index 4, obtains corresponding one-level and passes judgment on collection, supposes that the one-level judge collection that obtains to index 2 is S2=ω 2 * Y2=(0.5241,0.1717; 0.163,0.1477), the one-level judge collection that obtains to index 3 is S3=ω 3 * Y3=(0.1525,0.2475; 0.1525,0.4475), the one-level judge collection that obtains to index 4 is S4=ω 4 * Y4=(0.09,0.155; 0.3,0.455), the one-level of then confirming is passed judgment on matrix and is $S={\left(\begin{array}{cccc}0.0875& 0.155& 0.66& 0.0975\\ 0.5241& 0.1717& 0.163& 0.1477\\ 0.1525& 0.2475& 0.1525& 0.4475\\ 0.09& 0.155& 0.3& 0.455\end{array}\right)}_{4\×4}.$

Testing apparatus is confirmed the relative importance degree value between index 1, index 2, index 3, the index 4, and structure one-level fuzzy evaluation matrix, supposes one-level weight sets ω=(0.2333 of confirming according to the one-level fuzzy evaluation matrix of structure; 0.275,0.2333,0.2583); Then one-level weight sets ω and one-level judge matrix S are multiplied each other, the performance that obtains is passed judgment on collection and is T=ω * S=(0.2234,0.1812; 0.3119; 0.2853), it is greatest member that visible performance is passed judgment on the element of concentrating 0.3119, its corresponding performance rate is the tertiary gradient; Therefore with this tertiary gradient as the network equipment being tested resulting test result, this test result is the combination property of this network equipment of test.

In addition, be subdivided into three grades, level Four or more how timely, also can adopt said method upwards to confirm the combination property of the network equipment step by step for index.

Fig. 3 testing apparatus structural representation corresponding with method of testing shown in Figure 1 that provide for the embodiment of the invention specifically comprises:

Test module 301 is used for each index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each index;

Weight sets determination module 302; Be used for according to the corresponding a plurality of test values of each index that obtain; Confirm the relative importance degree value between each index; With the relative importance degree value between each index of confirming is element structure fuzzy evaluation matrix, confirms weight sets according to the fuzzy evaluation matrix of structure;

Pass judgment on subclass determination module 303; Be used for to each index; Based on the corresponding a plurality of test values of this index that obtain; And the corresponding index test value scope of setting of each performance rate; Confirming the percentage of this index on each performance rate of test, is that the element structure is passed judgment on subclass with the percentage of this index on each performance rate of testing;

Test result determination module 304; Be used for according to being directed against the judge subclass that each index is confirmed respectively; Confirm to pass judgment on matrix; With weight sets of confirming and judge matrix multiple, the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on the corresponding performance rate of element of concentrating maximum confirm as the test result that obtains.

Fig. 4 testing apparatus structural representation corresponding with another kind of method of testing shown in Figure 2 that provide for the embodiment of the invention specifically comprises:

Test module 401 is used for each two-level index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each two-level index;

Secondary weight sets determination module 402; Be used for each first class index to the said network equipment; According to the corresponding a plurality of test values of each two-level index under this first class index that obtains, confirm the relative significance level value between each two-level index under this first class index, be the corresponding secondary fuzzy evaluation matrix of this first class index of element structure with the relative significance level value between each two-level index of confirming; According to the secondary fuzzy evaluation matrix of structure, confirm the secondary weight sets that this first class index is corresponding;

Secondary is passed judgment on subclass determination module 403; Be used for to each two-level index under this first class index; According to this two-level index that obtains corresponding to individual test value; And the corresponding two-level index test value scope of setting of each performance rate; Confirming the percentage of this two-level index on each performance rate of test, is that the corresponding secondary of this two-level index of element structure is passed judgment on subclass with the percentage of this two-level index on each performance rate of testing;

One-level is passed judgment on collection determination module 404; Be used for according to passing judgment on subclass to the secondary of each the two-level index structure under this first class index respectively; Confirm the secondary judge matrix that this first class index is corresponding; The secondary that the secondary weight sets that this first class index is corresponding is corresponding with this first class index is passed judgment on matrix multiple, obtains the corresponding one-level of this first class index and passes judgment on collection;

One-level judge matrix determination module 405 is used for passing judgment on collection according to the one-level of confirming to each first class index of the said network equipment respectively, confirms one-level judge matrix;

Test result determination module 406; Be used for confirming the relative significance level value between each first class index of the said network equipment; With the relative significance level value between each first class index is element structure one-level fuzzy evaluation matrix; One-level fuzzy evaluation matrix according to structure is confirmed the one-level weight sets; One-level weight sets of confirming and one-level are passed judgment on matrix multiple, and the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

The embodiment of the invention provides a kind of method of testing and device; This method is repeatedly tested each index of the network equipment, obtains the corresponding a plurality of test values of each index, according to the structure of the relative importance degree value between each index fuzzy evaluation matrix; Fuzzy evaluation matrix according to structure is confirmed weight sets; And the definite percentage of each index on each performance rate of testing, confirm to pass judgment on matrix in view of the above, with weight sets of confirming and judge matrix multiple; The performance that obtains this network equipment is passed judgment on collection, and the wherein maximum corresponding performance rate of element is confirmed as the test result that obtains.Through said method, testing apparatus is all tested each index, and combines the influence of each index to the combination property of the network equipment, therefore can test accurately the combination property of the network equipment.

Obviously, those skilled in the art can carry out various changes and modification and the spirit and the scope that do not break away from the application to the application.Like this, belong within the scope of the application's claim and equivalent technologies thereof if these of the application are revised with modification, then the application also is intended to comprise these changes and modification interior.

Claims (14)

1. a method of testing is characterized in that, comprising:

Each index to the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each index;

Based on the corresponding a plurality of test values of each index that obtain; Confirm the relative importance degree value between each index; With the relative importance degree value between each index of confirming is element structure fuzzy evaluation matrix, confirms weight sets based on the fuzzy evaluation matrix of structure;

To each index; Based on the corresponding a plurality of test values of this index that obtain; And the corresponding index test value scope of setting of each performance rate; Confirming the percentage of this index on each performance rate of test, is that the element structure is passed judgment on subclass with the percentage of this index on each performance rate of testing;

According to the judge subclass of confirming to each index respectively; Confirm to pass judgment on matrix; With weight sets of confirming and judge matrix multiple, the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on the corresponding performance rate of element of concentrating maximum confirm as the test result that obtains.

2. the method for claim 1 is characterized in that, each index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each index, specifically comprises:

According to set point number, each index of the network equipment is carried out the test of set point number, obtain the corresponding a plurality of test values of each index;

According to the corresponding a plurality of test values of each index that obtain, confirm the relative importance degree value between each index, specifically comprise:

The corresponding a plurality of test values of each index to obtain are that element is constructed first test matrix:

$A={\left(\begin{array}{cccc}{x}_{11}& x_{12}& \·\·\·& x_{1m}\\ x_{21}& x_{22}& \·\·\·& x_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ x_{n1}& x_{n2}& \·\·\·& x_{\mathrm{nm}}\end{array}\right)}_{n\×m},$

Wherein, n is said set point number, and m is the number of the index of the said network equipment, x _IjBe the i time resulting test value of test that j index in m the index carried out;

Said first test matrix is carried out normalization handles, obtain second test matrix:

$A^{\′}={\left(\begin{array}{cccc}{x}_{11}^{\′}& x_{12}^{\′}& \·\·\·& x_{1m}^{\′}\\ x_{21}^{\′}& x_{22}^{\′}& \·\·\·& x_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ x_{n1}^{\′}& x_{n2}^{\′}& \·\·\·& x_{\mathrm{nm}}^{\′}\end{array}\right)}_{n\×m},$

Wherein, x ' _IjBe the x in first test matrix _IjNormalized value;

The transposed matrix and said second test matrix of said second test matrix are multiplied each other, obtain the 3rd test matrix:

$B=A^{\′T}\×A^{\′}={\left(\begin{array}{cccc}{\mathrm{\α}}_{11}& {\mathrm{\α}}_{12}& \·\·\·& {\mathrm{\α}}_{1m}\\ {\mathrm{\α}}_{21}& {\mathrm{\α}}_{22}& \·\·\·& {\mathrm{\α}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\α}}_{m1}& {\mathrm{\α}}_{m2}& \·\·\·& {\mathrm{\α}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

According to the fiducial value that is each target setting, each fiducial value is added in said the 3rd test matrix accordingly, obtain the 4th test matrix:

$B={\left(\begin{array}{cccc}{\mathrm{\α}}_{11}& {\mathrm{\α}}_{12}& \·\·\·& {\mathrm{\α}}_{1m}\\ {\mathrm{\α}}_{21}& {\mathrm{\α}}_{22}& \·\·\·& {\mathrm{\α}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\α}}_{m1}& {\mathrm{\α}}_{m2}& \·\·\·& {\mathrm{\α}}_{\mathrm{mm}}\\ {\mathrm{\β}}_{m+\mathrm{1,1}}& {\mathrm{\β}}_{m+\mathrm{1,2}}& \·\·\·& {\mathrm{\β}}_{m+1,m}\end{array}\right)}_{(m+1)\×m},$

Wherein, β _{M+1, i}Fiducial value for i index in m the index;

Said the 4th test matrix is carried out normalization handles, obtain the 5th test matrix:

$B^{\′}={\left(\begin{array}{cccc}{\mathrm{\α}}_{11}^{\′}& {\mathrm{\α}}_{12}^{\′}& \·\·\·& {\mathrm{\α}}_{1m}^{\′}\\ {\mathrm{\α}}_{21}^{\′}& {\mathrm{\α}}_{22}^{\′}& \·\·\·& {\mathrm{\α}}_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\α}}_{m1}^{\′}& {\mathrm{\α}}_{m2}^{\′}& \·\·\·& {\mathrm{\α}}_{\mathrm{mm}}^{\′}\\ {\mathrm{\β}}_{m+\mathrm{1,1}}^{\′}& {\mathrm{\β}}_{m+\mathrm{1,2}}^{\′}& \·\·\·& {\mathrm{\β}}_{m+1,m}^{\′}\end{array}\right)}_{(m+1)\×m},$

Wherein, α ' _IjBe the α in the 4th test matrix _IjNormalized value, β ' _{M+1, i}Be the β in the 4th test matrix _{M+1, i}Normalized value;

According to said the 5th test matrix, adopt following formula to confirm the distance of the fiducial value of each index and each index:

$d_{i,m+1}=\sqrt{\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{({\mathrm{\α}}_{\mathrm{ij}}^{\′}-{\mathrm{\β}}_{(m+1),j}^{\′})}^2},$

Wherein, i=1,2,3...m, d _{I, m+1}Distance for the fiducial value of i index in m the index and this i index;

To any two indexs, confirm the relative importance degree value between these two indexs based on following formula:

$r_{\mathrm{ij}}=\frac{d_{j,m+1}}{d_{i,m+1}+d_{j,m+1}},$

Wherein, r _IjBe the relative importance degree value of i index in m the index with respect to j index.

3. according to claim 1 or claim 2 method is characterized in that, is element structure fuzzy evaluation matrix with the relative importance degree value between each index of confirming, specifically comprises:

Construct following matrix as the fuzzy evaluation matrix:

$R={\left(\begin{array}{cccc}{r}_{11}& r_{12}& ...& r_{1m}\\ r_{21}& r_{22}& ...& r_{2m}\\ ...& ...& ...& ...\\ r_{m1}& r_{m2}& ...& r_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, r _IjBe the relative importance degree value of i index with respect to j index, r _IjWith r _JiWith value be 1, and r _IiValue be 0.5;

Fuzzy evaluation matrix based on structure is confirmed weight sets, specifically comprises:

Fuzzy evaluation matrix to structure carries out the fuzzy consensus processing, obtains Fuzzy consistent matrix M:

$M={\left(\begin{array}{cccc}{\mathrm{\ξ}}_{11}& {\mathrm{\ξ}}_{12}& ...& {\mathrm{\ξ}}_{1m}\\ {\mathrm{\ξ}}_{21}& {\mathrm{\ξ}}_{22}& ...& {\mathrm{\ξ}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{\ξ}}_{m1}& {\mathrm{\ξ}}_{m2}& ...& {\mathrm{\ξ}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, for the arbitrary element ξ among the Fuzzy consistent matrix M _Ij,

r _IkBe the element of the capable k row of i in the fuzzy evaluation matrix, r _JkElement for the capable k row of j in the fuzzy evaluation matrix;

According to Fuzzy consistent matrix M, adopt following formula to confirm the weight that each index is corresponding:

${\mathrm{\ω}}_i=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ξ}}_{\mathrm{ij}}-0.5}{\frac{m(m-1)}{2}},$

Wherein, ω _iIt is the corresponding weight of i index;

With the corresponding weight of confirming of each index is that element constitutes weight sets: ω=(ω ₁, ω ₂, ω ₃... ω _m).

4. the method for claim 1; It is characterized in that, according to the corresponding a plurality of test values of this index that obtain, and the corresponding index test value scope of setting of each performance rate; Confirm the percentage of this index on each performance rate of test, specifically comprise:

To each performance rate; In the corresponding a plurality of test values of this index of confirming to obtain; The number that belongs to the test value in the corresponding index test value scope of this performance rate; The ratio of the sum of a plurality of test values that the number of confirming is corresponding with this index that obtains is confirmed as the percentage of this index on this performance rate of test;

The percentage of this index on each performance rate with test is that the element structure is passed judgment on subclass, specifically comprises:

According to the percentage of this index on each performance rate, construct following judge subclass:

Y _i＝(y _i1，y _i2，y _i3...y _ip)，

Wherein, Y _iBe the judge subclass of i index, p is the number of performance rate, for Y _iIn arbitrary element y _Iq, y _IqBe the percentage of this i index on q performance rate.

5. method as claimed in claim 4 is characterized in that, based on the judge subclass of confirming to each index respectively, confirms to pass judgment on matrix, specifically comprises:

According to the judge subclass of confirming to each index respectively, confirm following judge matrix:

$Y={\left(\begin{array}{cccc}{y}_{11}& y_{12}& ...& y_{1p}\\ y_{21}& y_{22}& ...& y_{2p}\\ ...& ...& ...& ...\\ y_{m1}& y_{m2}& ...& y_{\mathrm{mp}}\end{array}\right)}_{m\×p},$

Wherein, for the arbitrary element y among the Y _Iq, y _IqBe the percentage of i index on q performance rate.

6. a method of testing is characterized in that, comprising:

Each two-level index to the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each two-level index;

Each first class index to the said network equipment; Based on the corresponding a plurality of test values of each two-level index under this first class index that obtains; Confirm the relative importance degree value between each two-level index under this first class index; With the relative importance degree value between each two-level index of confirming is the corresponding secondary fuzzy evaluation matrix of this first class index of element structure; Based on the secondary fuzzy evaluation matrix of structure, confirm the secondary weight sets that this first class index is corresponding;

To each two-level index under this first class index; Based on the corresponding a plurality of test values of this two-level index that obtain; And the corresponding two-level index test value scope of setting of each performance rate; Confirming the percentage of this two-level index on each performance rate of test, is that the corresponding secondary of this two-level index of element structure is passed judgment on subclass with the percentage of this two-level index on each performance rate of testing;

According to passing judgment on subclass to the secondary of each the two-level index structure under this first class index respectively; Confirm the secondary judge matrix that this first class index is corresponding; The secondary that the secondary weight sets that this first class index is corresponding is corresponding with this first class index is passed judgment on matrix multiple, obtains the corresponding one-level of this first class index and passes judgment on collection;

One-level according to confirming to each first class index of the said network equipment is respectively passed judgment on collection, confirms one-level judge matrix;

Confirm the relative importance degree value between each first class index of the said network equipment; With the relative importance degree value between each first class index is element structure one-level fuzzy evaluation matrix; One-level fuzzy evaluation matrix according to structure is confirmed the one-level weight sets; One-level weight sets of confirming and one-level are passed judgment on matrix multiple, and the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

7. method as claimed in claim 6 is characterized in that, each two-level index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each two-level index, specifically comprises:

According to set point number, each two-level index of the network equipment is carried out the test of set point number, obtain the corresponding a plurality of test values of each two-level index;

To each first class index of the said network equipment, according to the corresponding a plurality of test values of each two-level index under this first class index that obtains, confirm the relative significance level value between each two-level index under this first class index, specifically comprise:

To each first class index of the said network equipment, be that element is constructed first test matrix with the corresponding a plurality of test values of each two-level index under this first class index that obtains:

$\mathrm{Ag}={\left(\begin{array}{cccc}{\mathrm{xg}}_{11}& {\mathrm{xg}}_{12}& \·\·\·& {\mathrm{xg}}_{1m}\\ {\mathrm{xg}}_{21}& {\mathrm{xg}}_{22}& \·\·\·& {\mathrm{xg}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{xg}}_{n1}& {\mathrm{xg}}_{n2}& \·\·\·& {\mathrm{xg}}_{\mathrm{nm}}\end{array}\right)}_{n\×m},$

Wherein, n is said set point number, and m is the number of the two-level index under this first class index, xg _IjBe the i time resulting test value of test that j two-level index under g the first class index carried out;

Said first test matrix is carried out normalization handles, obtain second test matrix:

${\mathrm{Ag}}^{\′}={\left(\begin{array}{cccc}{\mathrm{xg}}_{11}^{\′}& {\mathrm{xg}}_{12}^{\′}& \·\·\·& {\mathrm{xg}}_{1m}^{\′}\\ {\mathrm{xg}}_{21}^{\′}& {\mathrm{xg}}_{22}^{\′}& \·\·\·& {\mathrm{xg}}_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{xg}}_{n1}^{\′}& {\mathrm{xg}}_{n2}^{\′}& \·\·\·& {\mathrm{xg}}_{\mathrm{nm}}^{\′}\end{array}\right)}_{n\×m},$

Wherein, xg ' _IjBe the xg in first test matrix _IjNormalized value;

The transposed matrix and said second test matrix of said second test matrix are multiplied each other, obtain the 3rd test matrix:

$\mathrm{Bg}={\mathrm{Ag}}^{\′T}\×{\mathrm{Ag}}^{\′}={\left(\begin{array}{cccc}{\mathrm{\αg}}_{11}& {\mathrm{\αg}}_{12}& \·\·\·& {\mathrm{\αg}}_{1m}\\ {\mathrm{\αg}}_{21}& {\mathrm{\αg}}_{22}& \·\·\·& {\mathrm{\αg}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\αg}}_{m1}& {\mathrm{\αg}}_{m2}& \·\·\·& {\mathrm{\αg}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

According to the fiducial value of setting for each two-level index under this first class index, each fiducial value is added in said the 3rd test matrix accordingly, obtain the 4th test matrix:

$\mathrm{Bg}={\left(\begin{array}{cccc}{\mathrm{\αg}}_{11}& {\mathrm{\αg}}_{12}& \·\·\·& {\mathrm{\αg}}_{1m}\\ {\mathrm{\αg}}_{21}& {\mathrm{\αg}}_{22}& \·\·\·& {\mathrm{\αg}}_{2m}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\αg}}_{m1}& {\mathrm{\αg}}_{m2}& \·\·\·& {\mathrm{\αg}}_{\mathrm{mm}}\\ {\mathrm{\βg}}_{m+\mathrm{1,1}}& {\mathrm{\βg}}_{m+\mathrm{1,2}}& \·\·\·& {\mathrm{\βg}}_{m+1,m}\end{array}\right)}_{(m+1)\×m},$

Wherein, β g _{M+1, i}Fiducial value for i two-level index under this g first class index; Said the 4th test matrix is carried out normalization handles, obtain the 5th test matrix:

${\mathrm{Bg}}^{\′}={\left(\begin{array}{cccc}{\mathrm{\αg}}_{11}^{\′}& {\mathrm{\αg}}_{12}^{\′}& \·\·\·& {\mathrm{\αg}}_{1m}^{\′}\\ {\mathrm{\αg}}_{21}^{\′}& {\mathrm{\αg}}_{22}^{\′}& \·\·\·& {\mathrm{\αg}}_{2m}^{\′}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{\αg}}_{m1}^{\′}& {\mathrm{\αg}}_{m2}^{\′}& \·\·\·& {\mathrm{\αg}}_{\mathrm{mm}}^{\′}\\ {\mathrm{\βg}}_{m+\mathrm{1,1}}^{\′}& {\mathrm{\βg}}_{m+\mathrm{1,2}}^{\′}& \·\·\·& {\mathrm{\βg}}_{m+1,m}^{\′}\end{array}\right)}_{(m+1)\×m},$

Wherein, α g ' _IjBe the α g in the 4th test matrix _IjNormalized value, β g ' _{M+1, i}Be the β g in the 4th test matrix _{M+1, i}Normalized value;

According to said the 5th test matrix, adopt following formula to confirm the distance of the fiducial value of each two-level index and each two-level index under this first class index:

${\mathrm{dg}}_{i,m+1}=\sqrt{\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{({\mathrm{\αg}}_{\mathrm{ij}}^{\′}-{\mathrm{\βg}}_{(m+1),j}^{\′})}^2},$

Wherein, i=1,2,3...m, dg _{I, m+1}Distance for the fiducial value of i two-level index under this g first class index and this i two-level index;

To any two two-level index under this first class index, confirm the relative importance degree value between these two two-level index based on following formula:

${\mathrm{rg}}_{\mathrm{ij}}=\frac{{\mathrm{dg}}_{j,m+1}}{{\mathrm{dg}}_{i,m+1}+{\mathrm{dg}}_{j,m+1}},$

Wherein, rg _IjBe the relative importance degree value of i two-level index under this g first class index with respect to j two-level index under this g first class index.

8. like claim 6 or 7 described methods, it is characterized in that, is the corresponding secondary fuzzy evaluation matrix of this first class index of element structure with the relative significance level value between each two-level index of confirming, specifically comprises:

Construct the secondary fuzzy evaluation matrix of following matrix as this first class index correspondence:

$\mathrm{Rg}={\left(\begin{array}{cccc}{\mathrm{rg}}_{11}& {\mathrm{rg}}_{12}& ...& {\mathrm{rg}}_{1m}\\ {\mathrm{rg}}_{21}& {\mathrm{rg}}_{22}& ...& {\mathrm{rg}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{rg}}_{m1}& {\mathrm{rg}}_{m2}& ...& {\mathrm{rg}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, rg _IjBe the relative importance degree value of i two-level index under this g first class index with respect to j two-level index under this g first class index, rg _IjWith rg _JiWith value be 1, and rg _IiValue be 0.5;

According to the secondary fuzzy evaluation matrix of structure, confirm the secondary weight sets that this first class index is corresponding, specifically comprise:

Secondary fuzzy evaluation matrix to structure carries out the fuzzy consensus processing, obtains secondary Fuzzy consistent matrix Mg:

$\mathrm{Mg}={\left(\begin{array}{cccc}{\mathrm{\ξg}}_{11}& {\mathrm{\ξg}}_{12}& ...& {\mathrm{\ξg}}_{1m}\\ {\mathrm{\ξg}}_{21}& {\mathrm{\ξg}}_{22}& ...& {\mathrm{\ξg}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{\ξg}}_{m1}& {\mathrm{\ξg}}_{m2}& ...& {\mathrm{\ξg}}_{\mathrm{mm}}\end{array}\right)}_{m\×m},$

Wherein, for the arbitrary element ξ g among the secondary Fuzzy consistent matrix Mg _Ij, Rg _IkBe the element of the capable k row of i in the secondary fuzzy evaluation matrix, rg _JkElement for the capable k row of j in the secondary fuzzy evaluation matrix;

According to secondary Fuzzy consistent matrix Mg, adopt following formula to confirm the corresponding weight of each two-level index under this first class index:

${\mathrm{\ωg}}_i=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ξg}}_{\mathrm{ij}}-0.5}{\frac{m(m-1)}{2}},$

Wherein, ω g _iBe the corresponding weight of i two-level index under this g first class index;

With the corresponding weight of each two-level index under this first class index of confirming is that element constitutes the corresponding secondary weight sets of this first class index: ω g=(ω g ₁, ω g ₂, ω g ₃... ω g _m).

9. method as claimed in claim 6; It is characterized in that, according to the corresponding a plurality of test values of this two-level index that obtain, and the corresponding two-level index test value scope of setting of each performance rate; Confirm the percentage of this two-level index on each performance rate of test, specifically comprise:

To each performance rate; In the corresponding a plurality of test values of this two-level index of confirming to obtain; The number that belongs to the test value in the corresponding index test value scope of this performance rate; The ratio of the sum of a plurality of test values that the number of confirming is corresponding with this two-level index that obtains is confirmed as the percentage of this two-level index on this performance rate of test;

The percentage of this two-level index on each performance rate with test is that the corresponding secondary of this two-level index of element structure is passed judgment on subclass, specifically comprises:

According to the percentage of this two-level index on each performance rate, construct following secondary and pass judgment on subclass:

Yg _i＝(yg _i1，yg _i2，yg _i3...yg _ip)，

Wherein, Yg _iThe secondary that is i two-level index under g the first class index is passed judgment on subclass, and p is the number of performance rate, for Yg _iIn arbitrary element yg _Iq, yg _IqBe the percentage of i two-level index on q performance rate under this g first class index.

10. method as claimed in claim 9 is characterized in that, based on passing judgment on subclass to the secondary of each the two-level index structure under this first class index respectively, confirms the secondary judge matrix that this first class index is corresponding, specifically comprises:

Secondary according to confirming to each two-level index under this first class index is respectively passed judgment on subclass, confirms following secondary judge matrix:

$\mathrm{Yg}={\left(\begin{array}{cccc}{\mathrm{yg}}_{11}& {\mathrm{yg}}_{12}& ...& {\mathrm{yg}}_{1p}\\ {\mathrm{yg}}_{21}& {\mathrm{yg}}_{22}& ...& {\mathrm{yg}}_{2p}\\ ...& ...& ...& ...\\ {\mathrm{yg}}_{m1}& {\mathrm{yg}}_{m2}& ...& {\mathrm{yg}}_{\mathrm{mp}}\end{array}\right)}_{m\×p},$

Wherein, for the arbitrary element yg among the Yg _Iq, yg _IqBe the percentage of i two-level index on q performance rate under this g first class index.

11. method as claimed in claim 10 is characterized in that, the secondary that the secondary weight sets that this first class index is corresponding is corresponding with this first class index is passed judgment on matrix multiple, obtains the corresponding one-level of this first class index and passes judgment on collection, specifically comprises:

Adopt following formula to confirm the one-level judge collection that this first class index is corresponding:

$\mathrm{Sg}=({\mathrm{\ωg}}_1,{\mathrm{\ωg}}_2,{\mathrm{\ωg}}_3...{\mathrm{\ωg}}_m)\×{\left(\begin{array}{cccc}{\mathrm{yg}}_{11}& {\mathrm{yg}}_{12}& ...& {\mathrm{yg}}_{1p}\\ {\mathrm{yg}}_{21}& {\mathrm{yg}}_{22}& ...& {\mathrm{yg}}_{2p}\\ ...& ...& ...& ...\\ {\mathrm{yg}}_{m1}& {\mathrm{yg}}_{m2}& ...& {\mathrm{yg}}_{\mathrm{mp}}\end{array}\right)}_{m\×p}=({\mathrm{sg}}_1,{\mathrm{sg}}_2,{\mathrm{sg}}_3...{\mathrm{sg}}_p),$

Wherein, Sg is that the corresponding one-level of g first class index is passed judgment on collection, ω g=(ω g ₁, ω g ₂, ω g ₃... ω g _m) be the corresponding secondary weight sets of confirming of this g first class index, sg ₁, sg ₂, sg ₃... sg _pJudge weight for this g the corresponding p of first class index difference performance rate;

One-level based on confirming to each first class index of the said network equipment is respectively passed judgment on collection, confirms one-level judge matrix, specifically comprises:

One-level based on confirming to each first class index of the said network equipment is respectively passed judgment on collection, confirms following one-level judge matrix:

$S={\left(\begin{array}{cccccc}{s1}_1& {s1}_2& ...& {s1}_q& ...& {s1}_p\\ {s2}_1& {s2}_2& ...& {s2}_q& ...& {s2}_p\\ ...& ...& ...& ...& ...& ...\\ {\mathrm{sg}}_1& {\mathrm{sg}}_2& ...& {\mathrm{sg}}_q& ...& {\mathrm{sg}}_p\\ ...& ...& ...& ...& ...& ...\\ {\mathrm{sh}}_1& {\mathrm{sh}}_2& ...& {\mathrm{sh}}_q& ...& {\mathrm{sh}}_p\end{array}\right)}_{h\×p},$

Wherein, S passes judgment on matrix for the one-level of confirming, h is the number of the first class index of the said network equipment, passes judgment on the arbitrary element sg in the matrix S for one-level _q, sg _qJudge weight for corresponding q the performance rate of g first class index in h the first class index of the said network equipment.

12. method as claimed in claim 6 is characterized in that, is element structure one-level fuzzy evaluation matrix with the relative significance level value between each first class index, specifically comprises:

Construct following matrix as said one-level fuzzy evaluation matrix:

$R={\left(\begin{array}{cccc}{r}_{11}& r_{12}& ...& r_{1h}\\ r_{21}& r_{22}& ...& r_{2h}\\ ...& ...& ...& ...\\ r_{h1}& r_{h2}& ...& r_{\mathrm{hh}}\end{array}\right)}_{h\×h},$

Wherein, h is the number of the first class index of the said network equipment, r _IjBe the relative importance degree value of i first class index in h the first class index of the said network equipment with respect to j first class index, r _IjWith r _JiWith value be 1, and r _IiValue be 0.5;

One-level fuzzy evaluation matrix according to structure is confirmed the one-level weight sets, specifically comprises:

One-level fuzzy evaluation matrix to structure carries out the fuzzy consensus processing, obtains one-level Fuzzy consistent matrix M:

$M={\left(\begin{array}{cccc}{\mathrm{\ξ}}_{11}& {\mathrm{\ξ}}_{12}& ...& {\mathrm{\ξ}}_{1h}\\ {\mathrm{\ξ}}_{21}& {\mathrm{\ξ}}_{22}& ...& {\mathrm{\ξ}}_{2h}\\ ...& ...& ...& ...\\ {\mathrm{\ξ}}_{h1}& {\mathrm{\ξ}}_{h2}& ...& {\mathrm{\ξ}}_{\mathrm{hh}}\end{array}\right)}_{h\×h},$

Wherein, for the arbitrary element ξ among the one-level Fuzzy consistent matrix M _Ij,

r _IkBe the element of the capable k row of i in the one-level fuzzy evaluation matrix, r _JkElement for the capable k row of j in the one-level fuzzy evaluation matrix;

According to one-level Fuzzy consistent matrix M, adopt following formula to confirm the weight that each first class index is corresponding:

${\mathrm{\ω}}_i=\frac{\underset{j=1}{\overset{h}{\mathrm{\Σ}}}{\mathrm{\ξ}}_{\mathrm{ij}}-0.5}{\frac{h(h-1)}{2}},$

Wherein, ω _iBe the corresponding weight of this i first class index;

With the corresponding weight of confirming of each first class index is that element constitutes the one-level weight sets: ω=(ω ₁, ω ₂, ω ₃... ω _h).

13. a testing apparatus is characterized in that, comprising:

Test module is used for each index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each index;

The weight sets determination module; Be used for according to the corresponding a plurality of test values of each index that obtain; Confirm the relative importance degree value between each index; With the relative importance degree value between each index of confirming is element structure fuzzy evaluation matrix, confirms weight sets according to the fuzzy evaluation matrix of structure;

Pass judgment on the subclass determination module; Be used for to each index; Based on the corresponding a plurality of test values of this index that obtain; And the corresponding index test value scope of setting of each performance rate; Confirming the percentage of this index on each performance rate of test, is that the element structure is passed judgment on subclass with the percentage of this index on each performance rate of testing;

The test result determination module; Be used for according to being directed against the judge subclass that each index is confirmed respectively; Confirm to pass judgment on matrix; With weight sets of confirming and judge matrix multiple, the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on the corresponding performance rate of element of concentrating maximum confirm as the test result that obtains.

14. a testing apparatus is characterized in that, comprising:

Test module is used for each two-level index of the network equipment is repeatedly tested, and obtains the corresponding a plurality of test values of each two-level index;

Secondary weight sets determination module; Be used for each first class index to the said network equipment; According to the corresponding a plurality of test values of each two-level index under this first class index that obtains, confirm the relative significance level value between each two-level index under this first class index, be the corresponding secondary fuzzy evaluation matrix of this first class index of element structure with the relative significance level value between each two-level index of confirming; According to the secondary fuzzy evaluation matrix of structure, confirm the secondary weight sets that this first class index is corresponding;

Secondary is passed judgment on the subclass determination module; Be used for to each two-level index under this first class index; According to this two-level index that obtains corresponding to individual test value; And the corresponding two-level index test value scope of setting of each performance rate; Confirming the percentage of this two-level index on each performance rate of test, is that the corresponding secondary of this two-level index of element structure is passed judgment on subclass with the percentage of this two-level index on each performance rate of testing;

One-level is passed judgment on the collection determination module; Be used for according to passing judgment on subclass to the secondary of each the two-level index structure under this first class index respectively; Confirm the secondary judge matrix that this first class index is corresponding; The secondary that the secondary weight sets that this first class index is corresponding is corresponding with this first class index is passed judgment on matrix multiple, obtains the corresponding one-level of this first class index and passes judgment on collection;

One-level judge matrix determination module is used for passing judgment on collection based on the one-level of confirming to each first class index of the said network equipment respectively, confirms one-level judge matrix;

The test result determination module; Be used for confirming the relative significance level value between each first class index of the said network equipment; With the relative significance level value between each first class index is element structure one-level fuzzy evaluation matrix; One-level fuzzy evaluation matrix according to structure is confirmed the one-level weight sets; One-level weight sets of confirming and one-level are passed judgment on matrix multiple, and the performance that obtains the said network equipment is passed judgment on collection, said performance is passed judgment on concentrated the corresponding performance rate of maximum element to confirm as the test result that obtains.

Images