CN104317853B - A kind of service cluster construction method based on Semantic Web - Google Patents

Abstract

The invention discloses a kind of service cluster construction method based on Semantic Web, comprise the following steps：S101, structure are based on semantic service clearance；S102, the mapping relations for setting up Web service and service clearance；S103, structure service cluster and its dynamic base；S104, logical Petri net are described to the institutional framework for servicing cluster.The all service clusters of present invention generation need to only quantify to n concept, then the number of times for searching body tree is n times, have large increase in the time complexity of generation service cluster；The present invention merges the species that specify that serve port and service quality by semantic concept, makes the structure of service cluster relatively reasonable；Present invention service cluster is generated in service clearance, the service cluster matching operation of Users ' Need-oriented is converted into the inquiry of coordinate, then service cluster construction basis are more, the structure precision for servicing cluster is higher, and system is smaller according to the time complexity of user's request matching service cluster, solves the limitation of general service clustering method.

Description

A kind of service cluster construction method based on Semantic Web

Technical field

The present invention relates to a kind of service cluster construction method based on Semantic Web.

Background technology

With the popularization of development with the application of Web service technology, quantity of service rapidly increases, different service provider's issues A large amount of functionally same or similar services, for service requester is provided compared with more options, while also increasing user Search and bind the difficulty for being adapted to the optimal service of oneself.Existing service mode is that user returns to a service for meeting request, There is trickle change or because the change of network environment causes currently to respond service failure when user's request, it is difficult to fast searching is replaced Generation service, the adaptivity of service response is poor, and become excessively complicated of the Services Composition process and model after combining is difficult to reality Existing dynamic adaptable.

In order to solve the above problems, some scholars propose the think of that service request and response are realized with the set of one group of service Think, and give bundle of services, service pool, the service concept such as cluster and service community (Service Community).Although they are each It is different, but essence is identical, is all before the service search for carrying out Users ' Need-oriented, first to carry out service cluster, by one group of work( Energy identical Services Aggregation together, service is provided as an entirety.

At present, several typical service clustering methods are occurred in that, such as Richi Nayak are by keyword in service description document The frequency of middle appearance is used as service similarity, it is proposed that the semantic description of extension Web service, and introduces Heterogeneous Web service point The thought of group, based on the above method, hierarchical clustering algorithm is applied to the cluster process of similar Web service again.For net Lattice are serviced, based on import of services output and function body similitude, it is proposed that a kind of service discovery method based on Ontology Clustering. In order to improve the cohesion of service cluster, Sun Ping etc. from similar two aspects of function phase Sihe process of service, service is carried out Cluster research.Meanwhile, on the basis of similarity is calculated, the clustering method based on the factor such as user's request and user experience It is suggested.

Similarity Measure is the main method of current cluster, but the factor for building cluster that different researchers chooses is Difference, wherein, correlation rule similarity, Words similarity, justice original similarity, parameter similarity and structural similarity are service phases Like the main calculation methods of degree.

Traditional clustering technique based on semantic similarity has limitation, is mainly reflected in the following aspects：

(1) the repeatedly concept based on body tree need to be carried out when, clustering to search, and typically requires parameter interface one in cluster Cause, body tree is integrated with substantial amounts of concept and construction is complicated, complicated the time required to multiple concept lookup is carried out in body tree Degree is larger；

(2) specified services classification is needed when, being polymerized, order is relatively low on classification of service, easily causes cluster incomplete The problems such as；

(3), the precision and service response of cluster are particularly thorny, integrated use service similarity calculating method, can be significantly Service Similarity Measure precision and the service discovery degree of accuracy are improved, however as the raising of Similarity Measure precision, classification of service Precision increases, and services the increasing number of class, then the service cluster matching primitives complexity increase of Users ' Need-oriented, service response effect Rate reduction.

The content of the invention

For above-mentioned technical problem present in prior art, the present invention proposes a kind of service cluster based on Semantic Web Construction method, solve the cluster time complexity occurred in general service clustering method it is high, classification the low problem of order, solution Contradiction between the precision and service response of service cluster of having determined, has merged parameter interface.

To achieve these goals, the present invention is adopted the following technical scheme that：

A kind of service cluster construction method based on Semantic Web, comprises the following steps：

S101, structure are based on semantic service clearance

Port atom Endpoint is defined, port atom is expressed as two tuples, Endpoint=(Description, Value)；

Wherein, Description is the description to port；Value is the quantized value of port atom, one end of unique mark Mouth atom, its value scope is real number field；

Mass atoms QoS is defined, mass atoms are expressed as a triple, QoS=(Description, Value, Values)；

Wherein, Description is the description to mass atoms；Value is the quantized value of mass atoms, and its value scope is Real number field；Values quantifies collection for the modification concept of mass atoms；

The generation flow of port atom and mass atoms is：

In Semantic Web, body E is given, after carrying out semantic tagger to Web service, by information extraction, obtain service Port concept set and quality concept collection, after based on semantic concept fusion, obtain port atom collection and mass atoms collection；

Service clearance Space is defined, if Space is the nonempty set of N-dimensional vector, N is port atom collection and mass atoms Total sum, F is a real number field, and Space multiplies closing for the addition and number of vector, i.e.,：If a, b ∈ Space, then a+b ∈Space；If a ∈ Space, c ∈ F, then c*a ∈ Space, then Space is called a service clearance, wherein：

Take that to determine N-dimensional complete 1 vectorial (1,1 ..1) be space coordinates base；

In the corresponding N-dimensional coordinate systems of service clearance Space, N number of reference axis is named as Endpoint ', QoS successively₁, QoS₂,…,QoS_n-1；

Whole Value values of port atom collection constitute mapping relations, N-1 mass atoms QoS with reference axis Endpoint ' Values values successively with reference axis QoS₁,QoS₂,…,QoS_n-1Mapping relations are constituted, mapping function is and multiplies 1 computing；

S102, the mapping relations for setting up Web service and service clearance

Web service Wservice is defined, Web service is expressed as seven tuples, Wservice=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R), wherein：

Descriptions is the various descriptions of Web service, including the description to quality；Id is the mark to servicing, can Uniquely determine a Web service；Endpoints represents the port atom collection of service；QoSs represents the mass atoms collection of service； Inputs is the |input paramete collection of service；Outputs is the output parameter set of service；R represent Endpoints and QoSs, The mapping relations of Inputs, Outputs and Descriptions and QoSs；

The Value values of mass atoms QoS are established rules then really in QoSs attributes：

By Wservice.R (Wservice.Descriptions, Wservice.QoSs), repairing for mass atoms QoS is obtained Decorations concept, according to locus of the concept in body tree, draws the quantized value of concept, and Value attributes to QoS are assigned Value；

Web service is as follows in the mapping ruler of service clearance：

If there is Web service Wservice₁=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs,R)；

(1), by Wservice₁.R(Wservice₁.Endpoints,Wservice₁.QoSs) understand, if Wservice₁.Endpoint_i∈Wservice₁.Endpoints with QoS_e,QoS_u,…,QoS_kCorrespondence, then set up set X= {Wservice₁.Endpoint_i,QoS_e,QoS_u,…,QoS_k, set up original coordinates vector V_i=(β₁,β₂,…,β_n), wherein right J=1 ..., n,

(2) if, Wservice₁.Endpoints={ Endpoint₁,Endpoint₂,…,Endpoint_m, m is nature Number, then set up m N-dimensional original coordinates vector V by formula (1)₁…V_m；

(3), original coordinates vector is converted to space coordinates, and transformation rule is：

If original coordinates vector is (β₁,β₂,…,β_n), then the space coordinates after converting is (γ₁,γ₂,…,γ_n), wherein J=1..n,

Even vector element is port atom, then be converted into the Value values of port atom；If vector element is former quality Son, then be converted into the Value values of mass atoms；If vector element is 0, constant；

Web service is set up by mapping ruler to be contacted with the mapping of service clearance, a Web service Wservice₁Energy Enough it is mapped as M space coordinates Z in service clearance₁…Z_m, wherein, M=| Wservice₁.Endpoints|；By service identifiers Wservice₁.Id as the M other mark Z of space coordinates₁(Id)…Z_m(Id) Z, is claimed₁(Id)…Z_m(Id) it is Web service Wservice₁Service atom；

Definition service atom Seratomic, service atom is expressed as two tuples, Seratomic=(Id, Coordinate)；Wherein：

Id is No. Id of Web service；Coordinate is mapped to the N-dimensional coordinate of service clearance coordinate system for Web service；

Definition service atomic distance, if in service clearance Space, there is two spaces coordinate A=(x₁,x₂,…x_n), B= (y₁,y₂,…,y_n), two service atom A (Id are mapped with coordinate A, B₁) and B (Id₂)；Service atomic distance H (A (Id₁), B(Id₂)) be：

S103, structure service cluster and its dynamic base

Definition service granularity, on the basis of in service clearance a bit, to the extension radius on each change in coordinate axis direction；

Definition service cluster Sercluster, service cluster is expressed as a triple, Sercluster=(Id, ρ, Seratomics), wherein：Id uniquely characterizes a service cluster；ρ is service granularity；Seratomics is service atom collection；

If presence service space S pace, it is ρ to give service granularity, then with each reference axis bounded range in coordinate system It is radius to service granularity ρ centered on arbitrary coordinate point, an axial symmetry space body Ω is formed in coordinate system；All mappings One service atom collection of service cluster of service atomic building in axial symmetry space body Ω, service atom collection is according to service atom Be the arrangement of keyword ascending order with the distance of centre coordinate, using the centre coordinate of Ω as service cluster mark Id；

Definition service cluster dynamic base Serclustdl, if presence service space S pace, it is ρ to give service granularity, then successively Centered on the whole coordinate points in each reference axis bounded range in coordinate system, to service granularity ρ as radius generation service cluster, The service cluster of all generations is summarized as an ordered set, and this ordered set is called the dynamic base for servicing cluster；

Service cluster dynamic base is expressed as a five-tuple, Serclustdl=(Version, ρ, Rule, Range, Serclusters)；Wherein：Version represents the version information of service cluster dynamic base；ρ is service granularity；Rule is service cluster The ordering rule of service gathering in dynamic base；Range is the border of service clearance reference axis；Serclusters is moved for service cluster Gathering is serviced in state storehouse to close；

S104, logical Petri net are described to the institutional framework for servicing cluster

(1), the logical Petri net description of service cluster dynamic base linkage update mechanism

If logical Petri net ∑₁=(P；T_D,T_I,T_O；F,I,O,M₀), wherein P={ p₁,p₂,p₃,p₄,p₅,p₆,p₇,p₈, p₉,p₁₀}；t₁∈T_I；f_I(t₁)=(p₁∨p₂)；t₂∈T_I；f_I(t₂)=(p₃∨p₄)；t₃∈T_I；f_I(t₃)=(p₅∨p₆∨p₇)； t₄∈T_I；f_I(t₄)=(p₈∨p₉)；

If dragging willing, p in place₁Representing serve port collection has renewal；p₂Representation quality atom collection has renewal；p₃Represent Service clearance has renewal；p₄Representing Web service description has renewal；p₅Representing service atom collection has renewal；p₆Representing service granularity has Update；p₇The Range attributes for representing service dynamic base have renewal；p₈The Rule attributes for representing service dynamic base are changed, that is, service Service gathering in dynamic base is closed ordering rule and is changed；p₉Representing service cluster has renewal；p₁₀Representing service cluster dynamic base has more Newly；t₁Expression is updated to service clearance；t₂Represent and service atom collection is updated；t₃Represent and service cluster be updated, t₄Represent and service cluster dynamic base is updated；

(2) the logical Petri net description that, service cluster is produced

If serve port collection has n element, there is m element, Web service k+1, service atom L+ in mass atoms collection 1, service cluster w+1, wherein n+m+1=q, q+k+1=v, v+L+3=z；

If ∑₂=(P；T_D,T_I,T_O；F,I,O,M₀), wherein P={ p₁,p₂,…,p_b+x+1}；t₁∈T_I；f_I(t₁)=((p₁∨ p₂∨…∨p_n)∧(p_n+1∨p_n+2∨…∨p_n+m∨.T.))；.T. logical truth is represented；t₂,t₃,...,t_k+5∈T_D；

p₁To p_nRepresent n serve port；p_n+1To p_n+mRepresent m mass atoms；p_qTo p_q+kRepresent k+1 Web service； p_vTo p_v+LRepresent L+1 service atom；p_v+L+1Represent service granularity；p_v+L+2Represent the coordinate set of bounded service clearance；p_zExtremely p_z+wRepresent w+1 service cluster；t₁Represent mapping Web service action；t₂To t_2+kExpression is serviced k+1 Web service respectively The coordinate mapping action in space；t_3+kRepresent and build service cluster action；t_4+kRepresent and build the dynamic base action of service cluster；

(3), the logical Petri net description of service cluster

Service cluster is a service atom collection, the function port association of service atom and Web service, function port with it is defeated Enter, export association, the logical Petri net model for servicing cluster is referred to as servicing cluster network element；

Definition service cluster network element is a logical Petri net ∑₃=(P；T_D,T_I,T_O；F,I,O,M₀), wherein：

P={ p₁,…,p_n+m+k,p_z,…,p_z+w}；t₁∈T_I&T_O；f_I(t₂)=(p₁∨p₂∨…∨p_n)；f_O(t₂)=(p_n+1 ∨p_n+2∨…∨p_n+m)；t₁∈T_I&T_O；f_I(t₁)=(p_n+1∨p_n+2∨…∨p_n+m)；f_O(t₁)=(p_z∨p_z+1∨…∨ p_z+w)；t₃∈T_I&T_O；f_I(t₃)=(p_n+1∨p_n+2∨…∨p_n+m)；f_O(t₃)=(p_n+m+1∨p_n+m+2∨…∨p_n+m+k)；

p₁To p_nRepresent n input；p_n+1To p_n+mRepresent m serve port；p_n+m+1To p_n+m+kRepresent k output；p_zExtremely p_z+wRepresent w+1 Web service；t₂Represent the trigger action of input and serve port；t₁Expression serve port reflects with Web service Penetrate action；t₃Represent the trigger action of serve port and output.

Further, in above-mentioned steps S101, the assignment method based on semantic concept fusion is：

For serve port concept set and quality concept collection, with the concept similarity computational methods based on Ontology；

According to locus of the concept in body tree, the quantized value of concept is drawn, and Value attributes to concept are carried out Mark；

Quantized value identical concept in concept set is omitted, result set is obtained；

For the either element that mass atoms are concentrated, the concept set of this element is further modified in collection body tree；

According to locus of the concept in body tree, the quantized value of concept is drawn；

Forming the Values attributes after quantifying to collect to mass atoms carries out assignment.

Further, it is necessary to judge the similitude of Web service, its decision method is in above-mentioned steps S103：

The similarity of definition service atom, is located in service clearance Space, there are two service atom A (Id₁) and B (Id₂), two similarities of service atom are service atomic distance H (A (Id₁),B(Id₂))；H=0, represents two service atoms It is definitely similar；

The similarity determination rule of definition service atom：If presence service space S pace, it is ρ to give service granularity, if The similarity of atom is serviced less than or equal to service granularity, then it is similar to service atom.

Further, the optimal similarity of Web service is defined：

If in the presence of two Web services, Wservice₁=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R) and Wservice₂=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R)； Wservice₁It is in the service atom collection of service clearance：P={ A₁(Id₁),A₂(Id₂),…,A_k(Id_k)},Wservice₂In clothes Be engaged in space service atom collection be：Q={ B₁(Id₁),B₂(Id₂),…,B_j(Id_j)}；

Then the optimal similarity X computing formula of Web service are：

Define the optimal similarity determination rule of Web service：

It is located in service clearance Space, it is ρ to give service granularity, there are two Web services, Wservice₁And Wservice₂；If less than or equal to service granularity, Web service is optimal similar to the optimal similarity of Web service.

Further, the average similarity of Web service is defined：

If in the presence of two Web services, Wservice₁And Wservice₂；Wservice₁In the service atom collection of service clearance For：P={ A₁(Id₁),A₂(Id₂),…,A_k(Id_k)},Wservice₂It is in the service atom collection of service clearance：Q={ B₁ (Id₁),B₂(Id₂),…,B_j(Id_j)}；

Then the average similarity X computing formula of Web service are：

Define the average similarity decision rule of Web service：

It is located in service clearance Space, it is ρ to give service granularity, there are two Web services, Wservice₁And Wservice₂；If the average similarity of Web service is less than or equal to service granularity, Web service is average similar.

The invention has the advantages that：

The service cluster construction method based on Semantic Web in the present invention, gives modified body tree, increases between concept Modified relationship is added, service clearance has been built on the basis of Ontology；Give the space reflection mechanism of Web service；Construct Service cluster and its dynamic base；Give the description of the logical Petri net of service cluster.The all service clusters of present invention generation only need to be to n Individual concept is quantified, then the number of times for searching body tree is n times, is had in the time complexity of generation service cluster and is carried greatly very much It is high；The present invention merges the species that specify that serve port and service quality by semantic concept, makes the structure of service cluster more Rationally；Present invention service cluster is generated in service clearance, and the service cluster matching operation of Users ' Need-oriented is converted into coordinate Inquiry, then service cluster construction basis it is more, service cluster structure precision it is higher, and system according to user's request matching service cluster Time complexity it is smaller, solve the limitation of general service clustering method.

Brief description of the drawings

Fig. 1 represents figure for the tree-shaped of body；

Fig. 2 is modified body tree table diagram；

Fig. 3 is the generation schematic flow sheet of middle port atom of the present invention, input atom and mass atoms；

Fig. 4 is the LPN model ∑s of service cluster dynamic base linkage update mechanism in the present invention₁Schematic diagram；

Fig. 5 be service in the present invention fasciation into LPN model ∑s₂Schematic diagram；

Fig. 6 is service cluster network element schematic diagram in the present invention；

Fig. 7 is the structure schematic diagram of body tree in emulation experiment of the present invention；

Fig. 8 is the space reflection schematic diagram of Web service in emulation experiment of the present invention；

Fig. 9 is the structure schematic diagram of service cluster in emulation experiment of the present invention.

Specific embodiment

Below in conjunction with the accompanying drawings and specific embodiment is described in further detail to the present invention：

A kind of service cluster construction method based on Semantic Web, comprises the following steps：

S101, structure are based on semantic service clearance

First, the concept of Semantic Web, body and Similarity Measure is introduced.

(1), Semantic Web (Semantic Web)

Proposed first in 1998 by the inventor Tim Berners-Lee of WWW." Semantic Web is to current Web Extension, and by increasing machine understandable semanteme in Web, preferably facilitate between machine and people and interoperated, so that Allow computer preferably and person cooperative work." build artificial intelligence field ontology (Ontology) Research foundation it On semantic network technology, can be set up to thing between the computers, between people and computer by semantic formal definition The uniformity of physical solution.

(2), body (Ontology)

Originate from philosophy field earliest, be defined as " explanation and explanation of a system of objective reality, it is extension One abstract entities ".It is an explanation for system to objective reality, does not rely on any specific language.The mesh of body Mark is the related notion in abstract certain field, determines the vocabulary of common accreditation in the field, and to the correlation between vocabulary, A formal definitions with different levels and explanation are provided, so as to set up a base being commonly understood by the domain knowledge Plinth.Semantic description based on Ontological concept is added by relevant information to Web service, by machine to the reason of all Web services Solution is set up on a platform for the common accreditation of agreement, so as to realize the operation to Web service semantic class.

Body is typically expressed as a directed tree, referred to as body tree (Ontology Tree).Node describes concept, node Between line correspond to concept between semantic relation.The tree-shaped that Fig. 1 gives body represents, wherein, Car and USA Car Between line illustrate inheritance between subclass and parent.

Fig. 2 gives a kind of improved body tree, increased the modified relationship between concept, wherein, Old, New and Car Between dotted line show that concept Old, New can modify concept Car.

(3), calculated based on semantic concept similarity

Two classes can be substantially divided into the research of semantic similarity between concept both at home and abroad：

One is to utilize semantic dictionary, such as the tree-like hierarchy system knot of the former composition of the synonym in WordNet, HowNet or justice Structure, by calculating comentropy or semantic distance between two concepts, calculates semantic similarity between concept；Two is to utilize corpus The method of statistics, according to the frequency that two concepts occur in upper and lower text, calculates semantic similarity between concept.

The definition of port atom, input atom and mass atoms is given below

Port atom Endpoint is defined, port atom is expressed as two tuples, Endpoint=(Description, Value)；

Wherein：Description is the description to port；Value is the quantized value of port atom, one end of unique mark Mouth atom, its value scope is real number field.

Mass atoms QoS is defined, mass atoms are expressed as a triple, QoS=(Description, Value, Values)；

Wherein：Description is the description to mass atoms；Value is the quantized value of mass atoms, and its value scope is Real number field；Values quantifies collection for the modification concept of mass atoms.

Port atom, the generation flow of mass atoms are：In Semantic Web, body E is given, semanteme is carried out to Web service After mark, by information extraction, port concept set, the quality concept collection of service are obtained.After based on semantic concept fusion, Obtain port atom collection and mass atoms collection.As shown in Figure 3.

It is based on semantic concept fusion assignment rule：

For port, quality concept collection, with the concept similarity computational methods based on Ontology.According to concept at this Locus in body tree, draws the quantized value of concept, and Value attributes to concept are labeled.Omit concept intensive quantity Change value identical concept, obtains result set.For the either element that mass atoms are concentrated, further this is modified in collection body tree The concept set of element.According to locus of the concept in body tree, the quantized value of concept is drawn.Formed after quantifying to collect to quality The Values attributes of atom carry out assignment.

The definition of service clearance is given below

Service clearance Space is defined, if Space is the nonempty set of N-dimensional vector, N is port atom collection and mass atoms Total sum, F is a real number field, and Space multiplies closing for the addition and number of vector, i.e.,：If a, b ∈ Space, then a+b ∈Space；If a ∈ Space, c ∈ F, then c*a ∈ Space, then Space is called a service clearance, wherein：

(1), take that to determine N-dimensional complete 1 vectorial (1,1 ..1) be space coordinates base；

(2), in the corresponding N-dimensional coordinate systems of service clearance Space, N number of reference axis is named as Endpoint ', QoS successively₁, QoS₂,…,QoS_n-1；

(3), whole Value values of port atom collection and reference axis Endpoint ' constitute mapping relations, and N-1 quality is former The Values values of sub- QoS successively with reference axis QoS₁,QoS₂,…,QoS_n-1Mapping relations are constituted, mapping function is and multiplies 1 computing.

Service clearance has merged port atom and mass atoms, abstract and quantified based on semantic service describing and about Beam.

S102, the mapping relations for setting up Web service and service clearance

Web service Wservice is defined, Web service is expressed as seven tuples,

Wservice=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R), wherein：

Descriptions is the various descriptions of Web service, including the description to quality；Id is the mark to servicing, can Uniquely determine a Web service；Endpoints represents the port atom collection of service；QoSs represents the mass atoms collection of service； Inputs is the |input paramete collection of service；Outputs is the output parameter set of service；R represent Endpoints and QoSs, The mapping relations of Inputs, Outputs and Descriptions and QoSs.

The Value values of mass atoms QoS are established rules really in QoSs attributes, are：

By Wservice.R (Wservice.Descriptions, Wservice.QoSs), repairing for mass atoms QoS is obtained Decorations concept, according to locus of the concept in body tree, draws the quantized value of concept, and Value attributes to QoS are carried out Assignment.

The mapping ruler that Web service is given below in service clearance is as follows：

If there is Web service Wservice₁=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs,R)；

(1), by Wservice₁.R(Wservice₁.Endpoints,Wservice₁.QoSs) understand, if Wservice₁.Endpoint_i∈Wservice₁.Endpoints with QoS_e,QoS_u,…,QoS_kCorrespondence, then set up set X= {Wservice₁.Endpoint_i,QoS_e,QoS_u,…,QoS_k, set up original coordinates vector V_i=(β₁,β₂,…,β_n), wherein right J=1 ..., n,

(2) if, Wservice₁.Endpoints={ Endpoint₁,Endpoint₂,…,Endpoint_m, m is nature Number, then set up m N-dimensional original coordinates vector V by formula (1)₁…V_m。

(3), original coordinates vector is converted to space coordinates, and transformation rule is：

If original coordinates vector is (β₁,β₂,…,β_n), then the space coordinates after converting is (γ₁,γ₂,…,γ_n), wherein J=1..n,

Even vector element is port atom, then be converted into the Value values of port atom；If vector element is former quality Son, then be converted into the Value values of mass atoms；If vector element is 0, constant.

For example, mark vector is (Endpoint ', 0, QoS₂,…,QoS_n-1), then space coordinates is (Endpoint ' .Value,0,QoS₂.Value,…,QoS_n-1.Value)。

Web service is set up by mapping ruler to be contacted with the mapping of service clearance, a Web service Wservice₁Energy Enough it is mapped as M space coordinates Z in service clearance₁…Z_m, wherein, M=| Wservice₁.Endpoints|；By service identifiers Wservice₁.Id as the M other mark Z of space coordinates₁(Id)…Z_m(Id) Z, is claimed₁(Id)…Z_m(Id) it is Web service Wservice₁Service atom.

Definition service atom Seratomic, service atom is expressed as two tuples, Seratomic=(Id, Coordinate)；Wherein：

Id is No. Id of Web service；Coordinate is mapped to the N-dimensional coordinate of service clearance coordinate system for Web service.

The concept of service atomic distance is introduced below

Definition service atomic distance, if in service clearance Space, there is two spaces coordinate A=(x₁,x₂,…x_n), B= (y₁,y₂,…,y_n), two service atom A (Id are mapped with coordinate A, B₁) and B (Id₂)；Service atomic distance H (A (Id₁), B(Id₂)) be：

S103, structure service cluster and its dynamic base

Definition service granularity, on the basis of in service clearance a bit, to the extension radius on each change in coordinate axis direction.

An axial symmetry space body of service clearance can be marked off by service granularity, its radius is service granularity.Below Provide the similarity determination method of Web service：

The similarity of definition service atom, is located in service clearance Space, there are two service atom A (Id₁) and B (Id₂), two similarities of service atom are service atomic distance H (A (Id₁),B(Id₂))；H=0, represents two service atoms It is definitely similar.

The similitude for servicing atom illustrates the similitude met customer need.Absolute similar representation service is interatomic Serve port and quality constraint are identical.

The similarity determination rule of definition service atom：

If presence service space S pace, it is ρ to give service granularity, if the similarity of service atom is less than or equal to service Granularity, then it is similar to service atom.

Above-mentioned definition gives the similarity determination rule of service atom, and a Web service can include some service ends Mouthful, comprising some service origin and its characteristic atomics.Similitude between analysis Web service, can be from maximum comparability, average phase Like property and local similarity aspect analysis.

Shown below is the optimal similarity definition of Web service：

The optimal similarity of Web service is defined, if in the presence of two Web services, Wservice₁=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R) and Wservice₂=(Descriptions, Id, Endpoints, QoSs,Inputs,Outputs,R)；Wservice₁It is in the service atom collection of service clearance：P={ A₁(Id₁),A₂ (Id₂),…,A_k(Id_k)},Wservice₂It is in the service atom collection of service clearance：Q={ B₁(Id₁),B₂(Id₂),…,B_j (Id_j)}；

Then the optimal similarity X computing formula of Web service are：

Define the optimal similarity determination rule of Web service：

It is located in service clearance Space, it is ρ to give service granularity, there are two Web services, Wservice₁And Wservice₂；If less than or equal to service granularity, Web service is optimal similar to the optimal similarity of Web service.

The average similarity definition of Web service is given below：

The average similarity of Web service is defined, if in the presence of two Web services, Wservice₁And Wservice₂； Wservice₁It is in the service atom collection of service clearance：P={ A₁(Id₁),A₂(Id₂),…,A_k(Id_k)},Wservice₂In clothes Be engaged in space service atom collection be：Q={ B₁(Id₁),B₂(Id₂),…,B_j(Id_j)}；

Then the average similarity X computing formula of Web service are：

Define the average similarity decision rule of Web service：

It is located in service clearance Space, it is ρ to give service granularity, there are two Web services, Wservice₁And Wservice₂；If the average similarity of Web service is less than or equal to service granularity, Web service is average similar.

The theory of Web service local similarity is to randomly select partial service atom according to system definition to carry out similarity meter Calculate and similarity determination.Its definition is similar with defined above.Service similarity determination is the precondition for building service cluster, below Provide definition and its construction method of service cluster.

Definition service cluster Sercluster, service cluster is expressed as a triple, Sercluster=(Id, ρ, Seratomics), wherein：Id uniquely characterizes a service cluster；ρ is service granularity；Seratomics is service atom collection.

Above-mentioned definition gives the basic structure of service cluster, and from above-mentioned definition, service cluster is a service atom Set, is given below construction method：

If presence service space S pace, it is ρ to give service granularity, then with each reference axis bounded range in coordinate system It is radius to service granularity ρ centered on arbitrary coordinate point, an axial symmetry space body Ω is formed in coordinate system；All mappings One service atom collection of service cluster of service atomic building in axial symmetry space body Ω, service atom collection is according to service atom Be the arrangement of keyword ascending order with the distance of centre coordinate, using the centre coordinate of Ω as service cluster mark Id.

From above-mentioned definition, a service cluster is one group of information organism of similar services atom, the clothes in service cluster Business atom has the similitude of similitude and the quality constraint of serve port, and the ability met customer need has similitude.Clothes The similitude of adjacent atom is maximum in business atom collection.Coordinate points in bounded service clearance are not unique, then the service cluster for producing has Several.

The definition of service cluster dynamic base is given below：

Definition service cluster dynamic base Serclustdl, if presence service space S pace, it is ρ to give service granularity, then successively Centered on the whole coordinate points in each reference axis bounded range in coordinate system, to service granularity ρ as radius generation service cluster, The service cluster of all generations is summarized as an ordered set, and this ordered set is called the dynamic base for servicing cluster；

Service cluster dynamic base is expressed as a five-tuple, Serclustdl=(Version, ρ, Rule, Range, Serclusters)；Wherein：Version represents the version information of service cluster dynamic base；ρ is service granularity；Rule is service cluster The ordering rule of service gathering in dynamic base；Range is the border of service clearance reference axis；Serclusters is moved for service cluster Gathering is serviced in state storehouse to close.

Above-mentioned definition gives service cluster dynamic library structure, and from above-mentioned definition, service cluster dynamic base is a service The ordered set of cluster.

S104, logical Petri net are described to the institutional framework for servicing cluster

(1), the logical Petri net description of service cluster dynamic base linkage update mechanism

If logical Petri net ∑₁=(P；T_D,T_I,T_O；F,I,O,M₀), as shown in figure 4, wherein：

P={ p₁,p₂,p₃,p₄,p₅,p₆,p₇,p₈,p₉,p₁₀}；t₁∈T_I；f_I(t₁)=(p₁∨p₂)；t₂∈T_I；f_I(t₂)= (p₃∨p₄)；t₃∈T_I；f_I(t₃)=(p₅∨p₆∨p₇)；t₄∈T_I；f_I(t₄)=(p₈∨p₉)。

To logical Petri net ∑₁It is described as follows：

If dragging willing, p in place₁Representing serve port collection has renewal；p₂Representation quality atom collection has renewal；p₃Represent Service clearance has renewal；p₄Representing Web service description has renewal；p₅Representing service atom collection has renewal；p₆Represent service granularity There is renewal；p₇The Range attributes for representing service dynamic base have renewal；p₈The Rule attributes for representing service dynamic base are changed, that is, take Service gathering in business dynamic base is closed ordering rule and is changed；p₉Representing service cluster has renewal；p₁₀Representing service cluster dynamic base has Update；t₁Expression is updated to service clearance；t₂Represent and service atom collection is updated；t₃Represent is carried out more to service cluster Newly, t₄Represent and service cluster dynamic base is updated.

(2) the logical Petri net description that, service cluster is produced

Assuming that：There is n element in serve port collection, mass atoms collection has m element, and Web service k+1, service is former Sub- L+1, service cluster w+1, wherein n+m+1=q, q+k+1=v, v+L+3=z.

If ∑₂=(P；T_D,T_I,T_O；F,I,O,M₀), as shown in figure 5, wherein P={ p₁,p₂,…,p_b+x+1}；t₁∈T_I；f_I (t₁)=((p₁∨p₂∨…∨p_n)∧(p_n+1∨p_n+2∨…∨p_n+m∨.T.))；t₂,t₃,...,t_k+5∈T_D。

p₁To p_nRepresent n serve port；p_n+1To p_n+mRepresent m mass atoms；p_qTo p_q+kRepresent k+1 Web service； p_vTo p_v+LRepresent L+1 service atom；p_v+L+1Represent service granularity；p_v+L+2Represent the coordinate set of bounded service clearance；p_zExtremely p_z+wRepresent w+1 service cluster；t₁Represent mapping Web service action；t₂To t_2+kExpression is serviced k+1 Web service respectively The coordinate mapping action in space；t_3+kRepresent and build service cluster action；t_4+kRepresent and build the dynamic base action of service cluster.

(3), the logical Petri net description of service cluster

Service cluster is a service atom collection.The function port association of service atom and Web service, function port with it is defeated Enter, export association.The logical Petri net model for servicing cluster is referred to as servicing cluster network element.

Definition service cluster network element is a logical Petri net ∑₃=(P；T_D,T_I,T_O；F,I,O,M₀), wherein：

P={ p₁,…,p_n+m+k,p_z,…,p_z+w}；t₁∈T_I&T_O；f_I(t₂)=(p₁∨p₂∨…∨p_n)；f_O(t₂)=(p_n+1 ∨p_n+2∨…∨p_n+m)；t₁∈T_I&T_O；f_I(t₁)=(p_n+1∨p_n+2∨…∨p_n+m)；f_O(t₁)=(p_z∨p_z+1∨…∨ p_z+w)；t₃∈T_I&T_O；f_I(t₃)=(p_n+1∨p_n+2∨…∨p_n+m)；f_O(t₃)=(p_n+m+1∨p_n+m+2∨…∨p_n+m+k)。

p₁To p_nRepresent n input；p_n+1To p_n+mRepresent m serve port；p_n+m+1To p_n+m+kRepresent k output；p_zExtremely p_z+wRepresent w+1 Web service；t₂Represent the trigger action of input and serve port；t₁Expression serve port reflects with Web service Penetrate action；t₃Represent the trigger action of serve port and output.

Shown below is the emulation experiment of the service cluster construction method based on Semantic Web：

In view of currently without relevant criterion platform and test data set, the present invention needs self-defined 150 according to experiment Web service, including service the information such as description, constraint.Service does not perform significant operation in itself, but these services exist The OWL-S specifications of extension are all followed on interface and quality description.From in terms of test angle, they do not have area with real Web service Not.

(1) body tree, is built

According to 150 the basic descriptions and constraint of Web service, self-defined body tree, as shown in Figure 7.It is general including quality Read：Expenses standard, response time；Including modification concept：Charge is high, it is low to charge, be free of charge, the response time is small, the response time is long Deng；Including port concept：Order train ticket, order bus ticket, weather forecast etc..For the sake of simplicity, concept is represented with letter.

(2), flow is produced to obtain port atom collection by port atom, mass atoms:Endpoints={ A, B, C, D, E }, Mass atoms collection:QoSs={ a, b }.A.value=1, B.value=2 are obtained by concept space distance in body tree, C.value=3, D.value=4, E.value=5, a.value=6, a.values={ 8,9,10,11,12 }, b.value =7, b.values={ 13,14,15,16,17 }.

(3) 3-dimensional service clearance Space, is built by port atom collection, mass atoms collection.

Scale { 1,2,3,4,5 } in the Value values { 1,2,3,4,5 } and reference axis Endpoint of port atom collection is one by one Correspondence.A pair of scale { 1,2,3,4,5 } in the values values { 8,9,10,11,12 } and reference axis Qos a of mass atoms a Should.A pair of scale { 1,2,3,4,5 } in the values values { 13,14,15,16,17 } and reference axis Qos b of mass atoms b Should.150 Web services are mapped to service clearance, generation services atom, as shown in Figure 8.

(4) service cluster, is built.It is radius to service granularity ρ=1 centered on using the coordinate (3,3,3) for randomly selecting, One space sphere Ω of generation, as shown in Figure 9.It is the clothes of (3,3,3) to be mapped to the service atomic building Id in space sphere Ω The service atom collection of business cluster, it is that keyword ascending order is arranged that service atom integrates according to service atom and the distance of Ω centre coordinates.Clothes It is 35 that atomicity is serviced in business cluster, and service cluster is (3,3,3) { (2,2,2) ..., (3,3,3) ..., (4,4,4) }.

(5), comparative analysis

The present invention proposes a kind of service clustering method based on service clearance.Port and qualitative character are generated by services set Collection, is quantified to service concept collection by space length of the concept in body tree, is merged, and built service clearance.By servicing To the mapping of service clearance, obtain with service cluster that space coordinates is mark.Service cluster has similar serve port and service Quality.User's request can be mapped as service clearance coordinate, and system services cluster according to user's request coordinate matching.

Compared to traditional clustering technique based on semantic similarity, the inventive method has three superiority：

(1) time complexity for, building service cluster is low, and the cluster of service is completed in service clearance：

Traditional service cluster need to carry out the concept based on body tree for several times and search, and typically require that parameter connects in cluster Mouth is consistent.Body tree is integrated with substantial amounts of concept, and construction is complicated, when concept for several times is carried out in body tree searching required Between complexity it is larger.Being provided with n concept needs Similarity Measure, then generate lookup 2 needed for a service clusterⁿIt is secondary.

The all service clusters of present invention generation need to only quantify to n concept, then the number of times for searching body tree is n times.Then There is large increase in the time complexity of generation service cluster.

(2), the structure of service cluster has reasonability：

One space coordinates, identifies specific serve port and service quality, i.e., specific service features.It is with coordinate The service cluster of mark, characterizes the set of the service atom the most similar to the service features of coordinate.Traditional service aggregating side Method is, it is necessary to the classification of specified services cluster.So in classification have randomness, easily cause classification it is unreasonable, cannot cluster completely Entirely.

The present invention merges the species that specify that serve port and service quality by semantic concept, make service cluster build compared with For reasonable.

(3), the structure of service cluster has feasibility：

The construction basis for servicing cluster can be description, quality, input, output, implementation process of service etc..Service the structure of cluster Build according to more, then the service cluster for building is more accurate, the service atom in service cluster is fewer, and user matches service in cluster is serviced Time complexity it is smaller.

The foundation that the present invention builds service cluster is serve port and service quality, and the proposition such as Sun Ping is based on work(under semanteme Can be with the service clustering method of process, the standard that description, input, output, implementation process for servicing etc. is all clustered as service. Each attribute of service, such as different parameters number, different implementation process, different service describings are so considered in classification Deng.Because the increase of forecast classification, the service number in service cluster are less, the increasing number of cluster is serviced, Users ' Need-oriented There is contradiction with service response in service cluster matching primitives complexity increase, the then precision for clustering.Present invention service cluster is in clothes Generated in business space, the service cluster matching operation of Users ' Need-oriented is converted into the inquiry of coordinate.Then service cluster construction basis It is more, the structure precision raising higher of cluster is serviced, and system is smaller according to the time complexity of user's request matching service cluster, solution Determine the limitation of general service clustering method.

Certainly, described above is only presently preferred embodiments of the present invention, and the present invention is not limited to enumerate above-described embodiment, should When explanation, any those of ordinary skill in the art are all equivalent substitutes for being made, bright under the teaching of this specification Aobvious variant, all falls within the essential scope of this specification, ought to be subject to protection of the invention.

Claims (5)

1. a kind of service cluster construction method based on Semantic Web, it is characterised in that comprise the following steps：

S101, structure are based on semantic service clearance

Port atom Endpoint is defined, port atom is expressed as two tuples, Endpoint=(Description, Value)；

Wherein, Description is the description to port；Value is the quantized value of port atom, and unique mark a port is former Son, its value scope is real number field；

Mass atoms QoS is defined, mass atoms are expressed as a triple, QoS=(Description, Value, Values)；

Wherein, Description is the description to mass atoms；Value is the quantized value of mass atoms, and its value scope is real number Domain；Values quantifies collection for the modification concept of mass atoms；

The generation flow of port atom and mass atoms is：

In Semantic Web, body E is given, after carrying out semantic tagger to Web service, by information extraction, obtain the port of service Concept set and quality concept collection, after based on semantic concept fusion, obtain port atom collection and mass atoms collection；

Service clearance Space is defined, if Space is the nonempty set of N-dimensional vector, N is port atom collection and mass atoms sum Sum, F is a real number field, and Space multiplies closing for the addition and number of vector, i.e.,：If a, b ∈ Space, then a+b ∈ Space；If a ∈ Space, c ∈ F, then c*a ∈ Space, then Space is called a service clearance, wherein：

Take that to determine N-dimensional complete 1 vectorial (1,1 ..1) be space coordinates base；

In the corresponding N-dimensional coordinate systems of service clearance Space, N number of reference axis is named as Endpoint ', QoS successively₁,QoS₂,…, QoS_n-1；

Whole Value values of port atom collection and reference axis Endpoint ' constitute mapping relations, N-1 mass atoms QoS's Values values successively with reference axis QoS₁,QoS₂,…,QoS_n-1Mapping relations are constituted, mapping function is and multiplies 1 computing；

S102, the mapping relations for setting up Web service and service clearance

Web service Wservice is defined, Web service is expressed as seven tuples,

Wservice=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R), wherein：

Descriptions is the various descriptions of Web service, including the description to quality；Id is the mark to servicing, can be unique Determine a Web service；Endpoints represents the port atom collection of service；QoSs represents the mass atoms collection of service；Inputs It is the |input paramete collection of service；Outputs is the output parameter set of service；R represent Endpoints and QoSs, Inputs, The mapping relations of Outputs and Descriptions and QoSs；

The Value values of mass atoms QoS are established rules then really in QoSs attributes：

By Wservice.R (Wservice.Descriptions, Wservice.QoSs), the modification for obtaining mass atoms QoS is general Read, according to locus of the concept in body tree, draw the quantized value of concept, and Value attributes to QoS carry out assignment；

Mapping ruler of the Web service in service clearance：

If there is Web service Wservice₁=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R)；

(1), by Wservice₁.R(Wservice₁.Endpoints,Wservice₁.QoSs) understand, if Wservice₁.Endpoint_i∈Wservice₁.Endpoints with QoS_e,QoS_u,…,QoS_kCorrespondence, then set up set X= {Wservice₁.Endpoint_i,QoS_e,QoS_u,…,QoS_k, set up original coordinates vector V_i=(β₁,β₂,…,β_n), wherein right J=1 ..., n,

(2) if, Wservice₁.Endpoints={ Endpoint₁,Endpoint₂,…,Endpoint_m, m is natural number, then M N-dimensional original coordinates vector V is set up by formula (1)₁…V_m；

(3), original coordinates vector is converted to space coordinates, and transformation rule is：

If original coordinates vector is (β₁,β₂,…,β_n), then the space coordinates after converting is (γ₁,γ₂,…,γ_n), wherein j= 1..n,

Even vector element is port atom, then be converted into the Value values of port atom；If vector element is mass atoms, It is converted into the Value values of mass atoms；If vector element is 0, constant；

Web service is set up by mapping ruler to be contacted with the mapping of service clearance, a Web service Wservice₁Can map It is M space coordinates Z in service clearance₁…Z_m, wherein, M=| Wservice₁.Endpoints|；By service identifiers Wservice₁.Id as the M other mark Z of space coordinates₁(Id)…Z_m(Id) Z, is claimed₁(Id)…Z_m(Id) it is Web service Wservice₁Service atom；

Definition service atom Seratomic, service atom is expressed as two tuples, Seratomic=(Id, Coordinate)；Wherein：

Id is No. Id of Web service；Coordinate is mapped to the N-dimensional coordinate of service clearance coordinate system for Web service；

Definition service atomic distance, if in service clearance Space, there is two spaces coordinate A=(x₁,x₂,…x_n), B=(y₁, y₂,…,y_n), two service atom A (Id are mapped with coordinate A, B₁) and B (Id₂)；Service atomic distance H (A (Id₁),B (Id₂)) be：

$H\left(A\right({\mathrm{Id}}_1),B({\mathrm{Id}}_2\left)\right)=\sqrt{{(x_1-y_1)}^2+{(x_2-y_2)}^2+\mathrm{...}+{(x_n-y_n)}^2}---\left(3\right)$

S103, structure service cluster and its dynamic base

Definition service granularity, on the basis of in service clearance a bit, to the extension radius on each change in coordinate axis direction；

Definition service cluster Sercluster, service cluster is expressed as a triple, Sercluster=(Id, ρ, Seratomics), wherein：Id uniquely characterizes a service cluster；ρ is service granularity；Seratomics is service atom collection；

If presence service space S pace, it is ρ to give service granularity, then with any in each reference axis bounded range in coordinate system It is radius to service granularity ρ centered on coordinate points, an axial symmetry space body Ω is formed in coordinate system；It is all to be mapped to axle One service atom collection of service cluster of service atomic building in symmetric space body Ω, service atom collection is according to service atom with The distance of heart coordinate is the arrangement of keyword ascending order, using the centre coordinate of Ω as service cluster mark Id；

Service cluster dynamic base Serclustdl is defined, if presence service space S pace, it is ρ to give service granularity, then successively with seat Centered on whole coordinate points in mark system in each reference axis bounded range, to service granularity ρ as radius generation service cluster, all The service cluster of generation is summarized as an ordered set, and this ordered set is called the dynamic base for servicing cluster；

Service cluster dynamic base is expressed as a five-tuple, Serclustdl=(Version, ρ, Rule, Range, Serclusters)；Wherein：Version represents the version information of service cluster dynamic base；ρ is service granularity；Rule is service cluster The ordering rule of service gathering in dynamic base；Range is the border of service clearance reference axis；Serclusters is moved for service cluster Gathering is serviced in state storehouse to close；

S104, logical Petri net are described to the institutional framework for servicing cluster

(1), the logical Petri net description of service cluster dynamic base linkage update mechanism

If logical Petri net ∑₁=(P；T_D,T_I,T_O；F,I,O,M₀), wherein P={ p₁,p₂,p₃,p₄,p₅,p₆,p₇,p₈,p₉, p₁₀}；t₁∈T_I；f_I(t₁)=(p₁∨p₂)；t₂∈T_I；f_I(t₂)=(p₃∨p₄)；t₃∈T_I；f_I(t₃)=(p₅∨p₆∨p₇)；t₄ ∈T_I；f_I(t₄)=(p₈∨p₉)；

If dragging willing, p in place₁Representing serve port collection has renewal；p₂Representation quality atom collection has renewal；p₃Represent service There is renewal in space；p₄Representing Web service description has renewal；p₅Representing service atom collection has renewal；p₆Representing service granularity has more Newly；p₇The Range attributes for representing service dynamic base have renewal；p₈The Rule attributes for representing service dynamic base are changed, that is, service dynamic Service gathering in state storehouse is closed ordering rule and is changed；p₉Representing service cluster has renewal；p₁₀Representing service cluster dynamic base has renewal； t₁Expression is updated to service clearance；t₂Represent and service atom collection is updated；t₃Represent and service cluster is updated, t₄ Represent and service cluster dynamic base is updated；

(2) the logical Petri net description that, service cluster is produced

If serve port collection has n element, there is m element in mass atoms collection, Web service k+1, service atom L+1 is individual, Service cluster w+1, wherein n+m+1=q, q+k+1=v, v+L+3=z；

If logical Petri net ∑₂=(P；T_D,T_I,T_O；F,I,O,M₀), wherein P={ p₁,p₂,…,p_b+x+1}；t₁∈T_I；f_I(t₁) =((p₁∨p₂∨…∨p_n)∧(p_n+1∨p_n+2∨…∨p_n+m∨.T.))；.T. logical truth is represented；t₂,t₃,...,t_k+5∈T_D；

p₁To p_nRepresent n serve port；p_n+1To p_n+mRepresent m mass atoms；p_qTo p_q+kRepresent k+1 Web service；p_vExtremely p_v+LRepresent L+1 service atom；p_v+L+1Represent service granularity；p_v+L+2Represent the coordinate set of bounded service clearance；p_zTo p_z+wGeneration W+1 service cluster of table；t₁Represent mapping Web service action；t₂To t_2+kExpression carries out service clearance to k+1 Web service respectively Coordinate mapping action；t_3+kRepresent and build service cluster action；t_4+kRepresent and build the dynamic base action of service cluster；

(3), the logical Petri net description of service cluster

Service cluster is a service atom collection, the function port association of service atom and Web service, function port be input into, it is defeated Go out association, the logical Petri net model for servicing cluster is referred to as servicing cluster network element；

Definition service cluster network element is a logical Petri net ∑₃=(P；T_D,T_I,T_O；F,I,O,M₀), wherein：

P={ p₁,…,p_n+m+k,p_z,…,p_z+w}；t₁∈T_I&T_O；f_I(t₂)=(p₁∨p₂∨…∨p_n)；f_O(t₂)=(p_n+1∨ p_n+2∨…∨p_n+m)；t₁∈T_I&T_O；f_I(t₁)=(p_n+1∨p_n+2∨…∨p_n+m)；f_O(t₁)=(p_z∨p_z+1∨…∨p_z+w)； t₃∈T_I&T_O；f_I(t₃)=(p_n+1∨p_n+2∨…∨p_n+m)；f_O(t₃)=(p_n+m+1∨p_n+m+2∨…∨p_n+m+k)；

p₁To p_nRepresent n input；p_n+1To p_n+mRepresent m serve port；p_n+m+1To p_n+m+kRepresent k output；p_zTo p_z+wGeneration W+1 Web service of table；t₂Represent the trigger action of input and serve port；t₁Represent that serve port is moved with the mapping of Web service Make；t₃Represent the trigger action of serve port and output.

2. a kind of service cluster construction method based on Semantic Web according to claim 1, it is characterised in that the step In S101, the assignment method based on semantic concept fusion is：

For serve port concept set and quality concept collection, with the concept similarity computational methods based on Ontology；

According to locus of the concept in body tree, the quantized value of concept is drawn, and Value attributes to concept enter rower Note；

Quantized value identical concept in concept set is omitted, result set is obtained；

For the either element that mass atoms are concentrated, the concept set of this element is further modified in collection body tree；

According to locus of the concept in body tree, the quantized value of concept is drawn；

Forming the Values attributes after quantifying to collect to mass atoms carries out assignment.

3. a kind of service cluster construction method based on Semantic Web according to claim 1, it is characterised in that the step , it is necessary to judge the similitude of Web service in S103, its decision method is：

The similarity of definition service atom, is located in service clearance Space, there are two service atom A (Id₁) and B (Id₂), two The similarity of individual service atom is service atomic distance H (A (Id₁),B(Id₂))；H=0, represents two service absolute phases of atom Seemingly；

The similarity determination rule of definition service atom：If presence service space S pace, it is ρ to give service granularity, if service Less than or equal to service granularity, then it is similar to service atom to the similarity of atom.

4. a kind of service cluster construction method based on Semantic Web according to claim 3, it is characterised in that

Define the optimal similarity of Web service：

If in the presence of two Web services, Wservice₁=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R) and Wservice₂=(Descriptions, Id, Endpoints, QoSs, Inputs, Outputs, R)； Wservice₁It is in the service atom collection of service clearance：P={ A₁(Id₁),A₂(Id₂),…,A_k(Id_k)},Wservice₂In clothes Be engaged in space service atom collection be：Q={ B₁(Id₁),B₂(Id₂),…,B_j(Id_j)}；

Then the optimal similarity X computing formula of Web service are：

$X=\underset{m=1}{\overset{k}{Min}}\left(\underset{i=1}{\overset{j}{Min}}\right(H\left(A_m\left({\mathrm{Id}}_m\right),B_i\left({\mathrm{Id}}_i\right)\right))---\left(4\right)$

Define the optimal similarity determination rule of Web service：

It is located in service clearance Space, it is ρ to give service granularity, there are two Web services, Wservice₁And Wservice₂； If less than or equal to service granularity, Web service is optimal similar to the optimal similarity of Web service.

5. a kind of service cluster construction method based on Semantic Web according to claim 3, it is characterised in that

Define the average similarity of Web service：

If in the presence of two Web services, Wservice₁And Wservice₂；Wservice₁It is in the service atom collection of service clearance：P ={ A₁(Id₁),A₂(Id₂),…,A_k(Id_k)},Wservice₂It is in the service atom collection of service clearance：Q={ B₁(Id₁),B₂ (Id₂),…,B_j(Id_j)}；

Then the average similarity X computing formula of Web service are：

$X=\left(\underset{m=1}{\overset{k}{\Σ}}\right(\underset{i=1}{\overset{j}{\Σ}}\left(H\left(A_m\left({\mathrm{Id}}_m\right),B_i\left({\mathrm{Id}}_i\right)\right)\right))/(k*j)---\left(5\right)$

Define the average similarity decision rule of Web service：

It is located in service clearance Space, it is ρ to give service granularity, there are two Web services, Wservice₁And Wservice₂； If the average similarity of Web service is less than or equal to service granularity, Web service is average similar.