CN103631925A - Fast grouping and retrieving method for machining equipment - Google Patents

Abstract

The invention relates to a fast grouping and retrieving method for machining equipment, and belongs to the field of machining and manufacturing. According to the method, the equipment is described by taking three dimensions of the type, the size and the machining precision of machining features as the machining feature vectors of manufacturing equipment; then, according to the feature vector description of the manufacturing equipment, the conventional extended fuzzy C-means algorithm for equipment grouping is improved and the machining feature vector-based equipment grouping algorithm is constructed; corresponding manufacturing equipment can be retrieved fast according to an optimal equipment grouping result. The fast grouping and retrieving for the machining equipment is an important link for manufacturability evaluation, process planning, production planning and the like. The efficiency of the algorithm is improved by disclosing a grouping method based on equipment machining feature vectors aiming at the current situation that the conventional fast grouping and retrieving algorithm for the equipment is low in efficiency.

Description

The fast grouping search method of machining equipment

Technical field

The fast grouping search method that the present invention relates to a kind of machining equipment, belongs to processing and manufacturing field.

Background technology

Under Networked Manufacturing, manufacturing equipment resource has the features such as kind is various, quantity is huge, working ability wide scope.These features select suitable process equipment to bring great difficulty to processing parts.And various this problem that makes of the complexity of processing parts is more outstanding.How in numerous optional equipments, selecting suitable process equipment with suitable processing cost, to complete the processing of part, is to improve the effective way that part is manufactured efficiency and cut down finished cost.The development of computer technology alleviates the workload of the equipment retrieval of knowing clearly to a certain extent, but manufacturing equipment systematic searching is needed in efficiency and method further improved, as required, first set grouping number, and can not solve grouping number optimization and the low inferior problem of recall precision, therefore also lack a kind of need not grouping in advance, and realize the method for device packets optimization and high efficiency selected, meet networking manufacture under demand to manufacturing equipment grouping and retrieval.

Fuzzy C-Means Clustering Algorithm can obtain the degree of uncertainty that sample belongs to each group, is applicable to manufacturing equipment grouping.Yet Fuzzy C-Means Clustering Algorithm requires first to set grouping number, can not solve the optimization problem of grouping number.Limitation for Fuzzy C-Means Clustering Algorithm, improved fuzzy C-clustering (Shi Xudong, Li Feng, Li Jun, Deng. towards the virtual manufacture Resource Modeling [J] of Manufacturability Evaluation. China Mechanical Engineering, 2002 (13): 1483-1485.), although the method has been considered the otherness between the consistance group in group simultaneously, but it is adopting traversal method aspect the optimization of group number, cause like this algorithm complex to increase, especially when number of devices is numerous, the efficiency of the method will be very low.Recent years some scholars algorithm that genetic algorithm combines with Fuzzy C-Means Clustering Algorithm that begins one's study, although can avoid Fuzzy C-Means Clustering Algorithm to be absorbed in local optimum.But, in the situation that packet count changes, the code length of gene can change, the probability of the infeasible solution that crisscross inheritance generates is too large, cause algorithm to be deep in and get rid of on infeasible solution, therefore the algorithm that, genetic algorithm combines with Fuzzy C-Means Clustering Algorithm is not suitable for solving the device packets problem based on machining feature.By researching and analysing above, the present invention adopts the fuzzy C-clustering of expansion to count optimization as manufacturing equipment group technology in conjunction with non-traversal group.

Summary of the invention

The object of the invention is in order to solve prior art grouping and the low problem of recall precision, the fast grouping search method of a kind of machining equipment disclosed by the invention, can improve machining equipment recall precision.

The object of the invention is to be achieved through the following technical solutions.

Machining equipment fast grouping search method, concrete steps are as follows:

Step 1, classification.

Taking into full account the process constraint information of machining feature and systematically analyzing on the basis of machining feature shape and project semantics information, for box parts, can summarize following machining feature and sort out principle:

1) from the viewpoint of processing and manufacturing rather than from describing the angle of design of part, machining feature is described;

2) should consider the hierarchical structure relation of feature;

3), according to the processing technology attribute of feature, define the definition on border between feature;

4) according to Manufacturability Evaluation, CAPP and the production schedule to parameters such as the size of the Location of requirement feature of feature, precision.

According to the classification principle of machining feature, the machining feature of box parts classifies as: face, post cone, groove, hole; Face comprises: logical plane, step surface, curved surface; Groove comprises: general groove, outer cutting annular groove, groove system; Hole comprises: general hole, hole system, threaded hole;

Step 2, description.

According to aspects such as the characteristic type of the classification of step 1 and machining feature, accessory size, machining feature accuracy requirements, the proper vector of manufacturing equipment can be described as on mathematics:

e _i＝(e _i,1,e _i,2,……,e _i,10,b _i,t _i)i＝1,2,……,n

Wherein:

Total number of n indication equipment;

E _i,krepresent that i equipment processes the feature capabilities numerical value of k feature.For example lathe is applicable to processing post cone feature, e _i,k=1; Lathe can be processed general hole characteristic, but not too applicable, e _i,k=0.5; Lathe cannot be processed logical plane characteristic, e _i,k=0.

By classification and the analysis to the feature description vectors of constitution equipment working ability of machining feature, under Networked Manufacturing, the feature description vectors of multi-processing equipment, can be expressed in matrix as:

Wherein, the corresponding relation that meets expression and apparatus characteristic description vectors of machining feature type is as shown in table 1.

The mapping table of table 1 machining feature and apparatus characteristic description vectors

Tab.1?Correspondence?table?of?manufacturing?feature?and?equipment?feature?matrix

Step 3, fuzzy grouping

1, initialization

Make initial packet group count m=1; Choose at random an equipment as initial packet center;

2, calculate the comprehensive evaluation index R of grouping

Calculate grouping group and count Clustering Comprehensive index R under m condition, R represents that average equipment irrelevance λ and machining feature repeat the sum reciprocal of index r.

$R=\mathrm{\λ}+\frac{1}{r}$

Wherein, average equipment irrelevance λ represents respectively to organize the mean value of equipment average departure degree:

$\mathrm{\λ}=\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\λ}}_j$

λ _jmanufacturing equipment M _ito group center G _j(G _jthe group center that represents j grouping) average departure degree.

R is that the machining feature of this grouping repeats index, by the degree of membership of machining feature, represents:

$r=\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\frac{1}{n_j}\underset{i=1}{\overset{n_j}{\mathrm{\Σ}}}{\left(g_{\mathrm{ij}}\right)}^2$

N _jthe quantity of the manufacturing equipment comprising in expression group j, the degree of membership g of machining feature k to group j _kjthe ratio that represents the weighted sum of apparatus characteristic description vectors in the weighted sum of the manufacturing equipment feature description vectors in each group and all groups with this machining feature.

$g_{\mathrm{kj}}=\frac{r_{\mathrm{kj}}}{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{r}_{\mathrm{kj}}}$

R _kjrepresent to have in j group the weighted sum of apparatus characteristic description vectors of the manufacturing equipment of machining feature k.

3, determine that optimum grouping group counts scope

According to step 32, can draw the scope of optimum grouping group number:

Wherein

while representing m=1, the irrelevance of equipment to group center;

4, upgrade grouping group and count m

Upgrade m=m+1, if

If grouping group is counted m=2, on Euclidean space, select two Euclidean distances apparatus characteristic description vectors farthest as first and second cluster centre G ₁and G ₂;

If grouping group is counted m>2:

Before note, the set at m-1 initial packet center is { G ₁..., G _m-1;

Calculate successively each apparatus characteristic description vectors { M _i| i=1,2 ..., n}(is the sum of n indication equipment wherein) and { G ₁..., G _m-1between distance B _i,s, wherein

{D _i,s＝‖M _i-G _i,k‖|s＝1,…,m-1}

And select equipment M _iminor increment min (D to front m-1 initial packet center _{i, 1}..., D _{i, m-1});

At all devices, in the minimum value at front m-1 initial packet center, select ultimate range, corresponding apparatus characteristic description vectors is taken as m group center G _m.

Min (D even _{i, 1}..., D _{i, m-1})=max{min{D _1,1..., D _{1, m-1}..., min{D _{n, 1}..., D _{n, m-1}, G _m=M _i;

Then repeating step 32 calculates grouping group and counts the comprehensive evaluation index R that m is corresponding;

Upgrade m=m+1, if select minimum Clustering Comprehensive index grouping as optimum grouping, optimum group result is comprised of three parts: the degree of membership of every group under optimum grouping group number, every group of manufacturing equipment comprising and each machining feature.

Step 4, retrieval

According to the characteristic type of machining feature, accessory size, machining feature accuracy requirement retrieval, can process the manufacturing equipment of this feature.First under in the optimum group result that Choice and process feature obtains in step 3, the group of degree of membership maximum is as the equipment group of this characteristic key, the hunting zone that the retrieval facility group of selecting is manufactured to resource as this characteristic key; Then in hunting zone, find out and can meet characteristic type, the accessory size of machining feature, the equipment of machining feature accuracy requirement; Finally extract the essential information of the equipment meeting the demands.

Principle of the present invention is feature description vectors e for manufacturing equipment _i=(e _{i, 1}, e _{i, 2}..., e _{i, 10}, b _i, t _i) describe, according to the principle of similarity maximum, group difference degree maximum in the group between apparatus characteristic description vectors, to the fuzzy grouping of manufacturing equipment, the present invention adopts the method validation of approximation of function to go out the scope of optimum grouping group number

this scope is really dwindled surely optimum grouping group and is counted hunting zone, improves the efficiency of device packets algorithm.Then according under feature not degree of membership size on the same group rapidly and efficiently retrieve corresponding manufacturing equipment.

Beneficial effect

The fast grouping search method of a kind of machining equipment of the present invention, to manufacturing equipment divides into groups and retrieval rapidly and accurately on the basis of describing in manufacturing equipment feature.The grouping problem of manufacturing equipment is that the equipment of similarity maximum is assigned to same group, and the equipment of diversity factor maximum is assigned to different groups, and the search problem of manufacturing equipment is according to degree of membership size deterministic retrieval scope and priority ranking.The present invention is describing on the basis of capacity of equipment with the proper vector of manufacturing equipment, has provided the method that Fuzzy C-Means Clustering Algorithm is determined optimum grouping group number in manufacturing equipment grouping of applying; Thereby improved recall precision, shortened retrieval time.

Accompanying drawing explanation

Fig. 1 is machining feature classification figure;

Fig. 2 is device packets process flow diagram.

Embodiment

Below in conjunction with accompanying drawing and embodiment, the present invention will be further described.

Embodiment 1

Machining equipment fast grouping search method, concrete steps are as follows:

Step 1, classification.

Machining feature is that geometry feature combines for processing the fundamental element of activity analysis with process constraint condition.Feature has been described the engineering significance of part geometry model.

Manufacture different its characteristic types that can process of resource kind also different, therefore can describe the working ability of manufacturing resource by the feature that manufacture resource can be processed.All machining feature that manufacture resource can be processed integrate just can form the machining feature description vectors of manufacturing resource.By being sorted out, machining feature can realize better the ability description of manufacturing resource.Taking into full account the process constraint information of machining feature and systematically analyzing on the basis of machining feature shape and project semantics information, for box parts, can summarize following machining feature and sort out principle:

5) from the viewpoint of processing and manufacturing rather than from describing the angle of design of part, machining feature is described;

6) should consider the hierarchical structure relation of feature;

7), according to the processing technology attribute of feature, define the definition on border between feature;

8) according to Manufacturability Evaluation, CAPP and the production schedule to parameters such as the size of the Location of requirement feature of feature, precision.

According to the classification principle of machining feature, the machining feature of box parts classifies as: face, post cone, groove, hole; Face comprises: logical plane, step surface, curved surface; Groove comprises: general groove, outer cutting annular groove, groove system; Hole comprises: general hole, hole system, threaded hole;

Step 2, description

In process, because size and the requirement on machining accuracy of part, machining feature are different, the feature of identical geometric type may need to use different manufacturing equipments to process.For example on part, the little hole of diameter can utilize drilling machine processing, but for the hole that diameter is large, precision is higher, just can not adopt drilling machine processing, also needs Boring machine processing hole; For another example, for general roughing or the semi-finishing of revolving parts cylindrical, need lathe to carry out turning, but may need cylindrical grinder to carry out grinding for the very high Excircle machining of accuracy requirement.Therefore, manufacturing equipment grouping is not only considered to want machinable characteristic type, also needs to divide into groups according to the size of part and the accuracy requirement etc. that whether meets machining feature.The feature of manufacturing equipment is described characteristic type, accessory size, the machining feature accuracy requirement three aspects: content that need to take into full account machining feature.

There are 50 of different types of machining tools in certain engineering shop, and every feature difference that lathe can be processed, is described as the proper vector of 50 manufacturing equipments in this workshop according to the classification of step 1:

e _i＝(e _i,1,e _i,2,……,e _i,10,b _i,t _i)i＝1,2,……,50

Wherein:

E _i,krepresent that i equipment processes the feature capabilities numerical value of k feature.For example lathe is applicable to processing post cone feature, e _i,k=1; Lathe can be processed general hole characteristic, but not too applicable, e _i,k=0.5; Lathe cannot be processed logical plane characteristic, e _i,k=0.

By classification and the analysis to the feature description vectors of constitution equipment working ability of machining feature, 50 manufacturing equipments in this workshop can be expressed as with matrix form:

Table 2 machine tooling feature description vectors table

Tab.2?List?of?MFV

Wherein, the symbol of the characteristic type of machining feature, accessory size, machining feature accuracy requirement represents with the corresponding relation of apparatus characteristic description vectors as shown in table 3.

Table 3 symbol represents the mapping table with apparatus characteristic description vectors

Tab.3?Correspondence?table?of?manufacturing?feature?and?equipment?feature?matrix

The equipment of take wherein illustrates its implication as example.For example be numbered 50 equipment, its feature description vectors { P, Pst, C, Ce, S, Sc, St, H, Hs, Ht, b, t} is { 1,0,0,0,1,0,0,0,0,0,0,1} represents to number 50 vertical machine and is applicable to the logical plane of processing, general cavity feature, can finishing feature, but can only process miniature parts;

Step 3, fuzzy grouping

1 initialization

Make initial packet group count m=1;

2 calculate corresponding initial packet center

If grouping group is counted m=1, random selected equipment 7 is as initial packet center;

If grouping group is counted m=2, on Euclidean space, select two Euclidean distances apparatus characteristic description vectors farthest as first and second cluster centre G ₁and G ₂, Euclidean distance equipment is farthest equipment 8 and 30 as calculated, the feature description vectors of equipment 8 and equipment 30 is respectively as first and second cluster centre G ₁and G ₂;

If grouping group is counted m>2:

When grouping group is counted m=3, calculate successively each apparatus characteristic description vectors { M _i| i=1,2 ..., n} and front 2 group center { G ₁, G ₂between distance B _{i, 1}and D _{i, 2}, in the minimum value of all devices feature description vectors and initial packet centre distance, select ultimate range,

D _i1=‖M _i-G ₁‖

D _i2=‖M _i-G ₂‖

And select equipment M _iminor increment min (D to front m-1 initial packet center _i1, D _i2);

At all devices, in the minimum value at front 2 initial packet centers, select ultimate range, corresponding apparatus characteristic description vectors is taken as the 3rd cluster centre G ₃, through calculating G ₃=M5, the feature description vectors of equipment 3 is as the 3rd initial packet center.

Calculate successively according to the method grouping group and count m=4, m=5, m=6, m=7, m=8, corresponding group center in m=9 situation.

3 calculate the comprehensive evaluation index R of grouping

Calculate grouping group and count Clustering Comprehensive index R under m condition, R represents that average equipment irrelevance λ and machining feature repeat the sum reciprocal of index r.

$R=\mathrm{\λ}+\frac{1}{r}$

Wherein, average equipment irrelevance λ represents respectively to organize the mean value of equipment average departure degree:

$\mathrm{\λ}=\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\λ}}_j$

λ _jmanufacturing equipment M _ito group center G _j(G _jthe group center that represents j grouping) average departure degree.

R is that the machining feature of this grouping repeats index, by the degree of membership of machining feature, represents:

$r=\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\frac{1}{n_j}\underset{i=1}{\overset{n_j}{\mathrm{\Σ}}}{\left(g_{\mathrm{ij}}\right)}^2$

N _jthe quantity of the manufacturing equipment comprising in expression group j, the degree of membership g of machining feature k to group j _kjthe ratio that represents the weighted sum of apparatus characteristic description vectors in the weighted sum of the manufacturing equipment feature description vectors in each group and all groups with this machining feature.

$g_{\mathrm{kj}}=\frac{r_{\mathrm{kj}}}{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{r}_{\mathrm{kj}}}$

R _kjrepresent to have in j group the weighted sum of apparatus characteristic description vectors of the manufacturing equipment of machining feature k.

Feature Description Matrix by 50 equipment of above-mentioned formula and this workshop can show that it is that { during 1,2,3,4,5,6,7,8,9}, the value of corresponding comprehensive evaluation index R is { 38,17.5,14,12.5,14.5,16,16.5,17,18} that grouping group is counted m.

4 definite optimum grouping groups are counted scope

Calculate the scope of optimum grouping group number:

wherein

while representing m=1, the irrelevance of equipment to group center.When calculating m=1, in the process of comprehensive evaluation index R, obtain

therefore, the scope of optimum grouping group number: m ∈ [1,9];

5 upgrade grouping group counts m

Upgrade m=m+1, if m is ∈ [1,9], 2 and 3 under repeating step three; Otherwise, select the grouping of minimum Clustering Comprehensive index to divide into groups as optimum.Because the fuzzy grouping effect maximum, group difference maximum of similarity in group is best, therefore comprehensive evaluation index R is the smaller the better, selecting grouping group number is [1,9] group result that conduct is optimum of corresponding comprehensive evaluation index R value minimum in scope, optimum group result is comprised of three parts: the degree of membership of every group under optimum grouping group number, every group of manufacturing equipment comprising and each machining feature.According to the R in example 1 step 3, can find out, when grouping group number is 4, comprehensive evaluation index is optimum, and corresponding optimum group result is as shown in table 4

Table 4 group result table

Tab.4?List?of?grouped?result

Take P (1), B (0.5) in first group, t (0.5) explains the implication of every stack features and degree of membership thereof as example, P (1) represents that logical plane characteristic is under the jurisdiction of the probability of first group and is approximately 100%, b (0.5) represents that the equipment in first group is applicable to semi-finishing, and t (0.5) represents that the equipment in first group is applicable to the medium-sized part of processing.

Step 4, retrieval

The optimum obtaining according to step 3, retrieves corresponding process equipment

For example, while requiring semi-finished step surface need to retrieve suitable manufacturing equipment on miniature parts, according to group result, we can see that Pst in second group (1), b (0.5), t (0.2) just in time satisfy condition, wherein described in Pst (1) expression step surface feature, the degree of membership of second group is 1, the lathe of all energy machine table terraces is nearly all in second group, b (0.5) represents that the equipment in second group is applicable to semi-finishing, and t (0.2) represents that the equipment of second group is applicable to the middle-size and small-size part of processing.Therefore, only need in second group, retrieve corresponding equipment, reduce significantly hunting zone, improve search efficiency.

Use MATLAB 7.0 on Intel Core i3-2310M platform, to realize checking high efficiency of the present invention.Utilize the fast grouping search method of machining equipment in traditional fuzzy C-clustering, improved FCM method and the present invention respectively the equipment in example to be carried out to packet authentication.Result of calculation is as shown in table 2.

Table 5 device packets method contrast table

Tab.5?comparison?table?of?different?grouping?algorithm

Analysis by the result shown in his-and-hers watches 5 is known:

(1) in improved FCM method and the present invention, the Clustering Comprehensive index R of method is less than traditional traditional fuzzy C-means clustering method, illustrates that the Clustering Effect in improved FCM method and the present invention is better than traditional Clustering Effect.

(2) in table 2, adopt method of the present invention and improved FCM method, there is identical Clustering Comprehensive index R value, but method 1.7s computing time of the present invention is few more a lot of than improved FCM method 15.6s computing time, illustrate that the inventive method efficiency is apparently higher than the efficiency of improved FCM method.

Above-described embodiment is described the preferred embodiment of the present invention; but design concept of the present invention is not limited to this; the unsubstantiality that those skilled in the art make technical scheme of the present invention improves, and all should fall in the definite protection domain of the claims in the present invention book.

Claims (1)

1. machining equipment fast grouping search method, is characterized in that: concrete steps are as follows:

Step 1, classification;

Taking into full account the process constraint information of machining feature and systematically analyzing on the basis of machining feature shape and project semantics information, for box parts, can summarize following machining feature and sort out principle:

1) from the viewpoint of processing and manufacturing rather than from describing the angle of design of part, machining feature is described;

2) should consider the hierarchical structure relation of feature;

3), according to the processing technology attribute of feature, define the definition on border between feature;

4) according to Manufacturability Evaluation, CAPP and the production schedule to parameters such as the size of the Location of requirement feature of feature, precision;

According to the classification principle of machining feature, the machining feature of box parts classifies as: face, post cone, groove, hole; Face comprises: logical plane, step surface, curved surface; Groove comprises: general groove, outer cutting annular groove, groove system; Hole comprises: general hole, hole system, threaded hole;

Step 2, description;

According to aspects such as the characteristic type of the classification of step 1 and machining feature, accessory size, machining feature accuracy requirements, the proper vector of manufacturing equipment can be described as on mathematics:

e _i＝(e _i,1,e _i,2,……,e _i,10,b _i,t _i)i＝1,2,……,n

Wherein:

Total number of n indication equipment;

E _i,krepresent that i equipment processes the feature capabilities numerical value of k feature; For example lathe is applicable to processing post cone feature, e _i,k=1; Lathe can be processed general hole characteristic, but not too applicable, e _i,k=0.5; Lathe cannot be processed logical plane characteristic, e _i,k=0;

By classification and the analysis to the feature description vectors of constitution equipment working ability of machining feature, under Networked Manufacturing, the feature description vectors of multi-processing equipment, can be expressed in matrix as:

Wherein, the corresponding relation that meets expression and apparatus characteristic description vectors of machining feature type is as shown in table 1;

The mapping table of table 1 machining feature and apparatus characteristic description vectors

Tab.1?Correspondence?table?of?manufacturing?feature?and?equipment?feature?matrix

Step 3, fuzzy grouping

1, initialization

Make initial packet group count m=1; Choose at random an equipment as initial packet center;

2, calculate the comprehensive evaluation index R of grouping

Calculate grouping group and count Clustering Comprehensive index R under m condition, R represents that average equipment irrelevance λ and machining feature repeat the sum reciprocal of index r;

$R=\mathrm{\λ}+\frac{1}{r}$

Wherein, average equipment irrelevance λ represents respectively to organize the mean value of equipment average departure degree:

$\mathrm{\λ}=\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\λ}}_j$

λ _jmanufacturing equipment M _ito group center G _j(G _jthe group center that represents j grouping) average departure degree;

R is that the machining feature of this grouping repeats index, by the degree of membership of machining feature, represents:

$r=\frac{1}{m}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\frac{1}{n_j}\underset{i=1}{\overset{n_j}{\mathrm{\Σ}}}{\left(g_{\mathrm{ij}}\right)}^2$

N _jthe quantity of the manufacturing equipment comprising in expression group j, the degree of membership g of machining feature k to group j _kjthe ratio that represents the weighted sum of apparatus characteristic description vectors in the weighted sum of the manufacturing equipment feature description vectors in each group and all groups with this machining feature;

$g_{\mathrm{kj}}=\frac{r_{\mathrm{kj}}}{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{r}_{\mathrm{kj}}}$

R _kjrepresent to have in j group the weighted sum of apparatus characteristic description vectors of the manufacturing equipment of machining feature k;

3, determine that optimum grouping group counts scope

According to step 32, can draw the scope of optimum grouping group number:

Wherein while representing m=1, the irrelevance of equipment to group center;

4, upgrade grouping group and count m

Upgrade m=m+1, if

If grouping group is counted m=2, on Euclidean space, select two Euclidean distances apparatus characteristic description vectors farthest as first and second cluster centre G ₁and G ₂;

If grouping group is counted m>2:

Before note, the set at m-1 initial packet center is { G ₁..., G _m-1;

Calculate successively each apparatus characteristic description vectors { M _i| i=1,2 ..., n}(is the sum of n indication equipment wherein) and { G ₁..., G _m-1between distance B _i,s, wherein

{D _i,s＝‖M _i-G _i,k‖|s＝1,…,m-1}

And select equipment M _iminor increment min (D to front m-1 initial packet center _{i, 1}..., D _{i, m-1});

At all devices, in the minimum value at front m-1 initial packet center, select ultimate range, corresponding apparatus characteristic description vectors is taken as m group center G _m;

Min (D even _{i, 1}..., D _{i, m-1})=max{min{D _1,1..., D _{1, m-1}..., min{D _{n, 1}..., D _{n, m-1}, G _m=M _i;

Then repeating step 32 calculates grouping group and counts the comprehensive evaluation index R that m is corresponding;

Upgrade m=m+1, if

select minimum Clustering Comprehensive index grouping as optimum grouping, optimum group result is comprised of three parts: the degree of membership of every group under optimum grouping group number, every group of manufacturing equipment comprising and each machining feature;

Step 4, retrieval

According to the characteristic type of machining feature, accessory size, machining feature accuracy requirement retrieval, can process the manufacturing equipment of this feature; First under in the optimum group result that Choice and process feature obtains in step 3, the group of degree of membership maximum is as the equipment group of this characteristic key, the hunting zone that the retrieval facility group of selecting is manufactured to resource as this characteristic key; Then in hunting zone, find out and can meet characteristic type, the accessory size of machining feature, the equipment of machining feature accuracy requirement; Finally extract the essential information of the equipment meeting the demands.

Images