CN102004767A - Abstract service logic-based interactive semantic Web service dynamic combination method - Google Patents

Abstract

The invention discloses an abstract service logic-based interactive semantic Web service dynamic combination method. The method comprises the following steps of: separating abstract service description from specific semantic Web service instances by establishing a resource service mapping model, performing unified semantic description on Web services by using an ontology web language for services (OWL-S), realizing static binding with lower-layer services, and providing a unified interface for the service combination; performing abstract description on the combined process through combined service process modeling and providing a combined template; finishing intelligent search and match of the services by combining Chinese word segmentation technology with an ontology-based conception matching algorithm under the support of a domain ontology and a professional thesaurus; and analyzing a timing sequence relationship and a control relationship of nodes in the combined process to realize dynamic binding and executing of the services, and constructing a service combination engine.

Description

A kind of interactive Semantic Web Services dynamic combined method based on abstract service logic

Technical field

The present invention relates to a kind of interactive Semantic Web Services dynamic combined method based on abstract service logic, promptly under the semantic Web service combination framework based on abstract operation flow, man-machine coordination dynamic creation operation flow is to realize integration of information system and resource sharing.This invention belongs to distributed manufacturing system information integration technical field.

Background technology

In the face of the challenge of kownledge economy and manufacturing globalization, service-oriented networking manufactures the development strategy for enterprise-essential in recent years.Brave academician of uncle Li Chinese Academy of Engineering etc. has proposed a kind of networking based on the cloud computing service mode in January, 2010 and has made new model---and the cloud manufacturing (Cloud manufacturing, CMfg).The cloud manufacturing is mainly towards manufacturing industry, enterprise product made required manufacturing resource consolidation become cloud to make service centre, and that the user can obtain as required at any time is safe and reliable, high-quality, all kinds of resource service cheaply.

And manufacture and design in the process at service-oriented product collaborative, each department realizes its core business by the Web service on the Internet.In order to improve the reusability of Web service, basic service can not be very complicated.Single simple in structure, the function singleness of Web service in the middle of actual often can't satisfy actual application demand, therefore usually needs to make up a plurality of basic Web services to finish complicated task.It will be more dynamically changeable that following enterprise product manufactures and designs in the process service environment, the various function services how effective combination distribution is developed by different tissues in network, realize seamless integrated between the service, form powerful operation flow service to realize integration of information system and resource sharing, become a major issue in manufacturing informatization field.

On synthetic framework of service and strategy, industry member tends to adopt the workflow modeling service procedure at present.Academia then tends to add semantic information into Web service, has proposed the service combining method based on AI (Artificial Intelligence) planning, and purpose is to realize the service Automatic Combined.Current service combining method based on workflow mainly is static combination, as BPEL.BPEL is the industry member main flow descriptive language of current Web Services Combination, directly adopting BPEL to carry out the Web service combination is a kind of method of artificial combination completely, the description that is abstract operation flow realizes binding in advance with concrete Web service, and basic service information in the flow process and execution result all are in advance as can be known.This method lacks automatism and dynamic, and is dumb and be difficult to safeguard, can't adapt to the system applies that loose lotus root is closed.At present based on the Web service automatic combination method of AI all based on certain formalization method or inference system, before combination, need carry out pre-service and formalization conversion to service.The complexity of these class methods can significantly improve greatly along with the change of planning space, also is in theoretical research stage at present.Will set up AI planning and Web service in the middle of actual and make up between the two mapping, also have a lot of problems to wait solution, that is that all right at present is ripe by computer realization Web service Automatic Combined fully.

Therefore, present stage realizes that the Web service Automatic Combined also must provide the process model of composite services to describe by the user.Manufacture and design in the process at enterprise product, the relative clear and definite of the task node in the operation flow makes that objectively the service synthesis system can the decomposition and combination service, obtains the array mode of its sub-services.The team that relates to numerous special disciplines and difference in functionality simultaneously at actual central product development process, it is impossible that professional's experience and knowledge all is converted to the intelligible formalized description of computing machine, finishes the wisdom that powerful collaborative design flow services musts be guarded or looked after all the time.Therefore, under the semantic Web service combination framework, based on abstract operation flow, by introducing the semantic service intellectuality that makes, thereby mode composite services with man-machine coordination, can satisfy users ' individualized requirement, improve the success ratio and the efficient of Services Combination, be actual at present a kind of service combining method.

Summary of the invention

(1) purpose: the present invention relates to a kind of interactive Semantic Web Services dynamic combined method based on abstract service logic, promptly under semantic Web service combination framework based on abstract operation flow, combine by experimental knowledge the intelligent and professional of computing machine, the establishment of combination process is finished in man-machine coordination, solving the uncertainty that exists in the AI planning and the problem of complexity problem and quiescent operation stream bad adaptability, thereby improve the success ratio and the efficient of Services Combination.

(2) technical scheme: related gordian technique comprises operation flow modeling, Semantic Web Services registration center, operation flow coupling, service semantics coupling, service search strategy and service dynamic binding in the Services Combination process.

The operation flow modeling is exactly to describe sequential relationship and control relation between the business activity with manual mode by patterned form.The user can create current non-existent, an abstract operation flow with new function according to actual needs.Modelling verification is by getting final product the description document of generation model self afterwards.The abstract operation flow of setting up can provide the function combinations template based on Web service, thereby improves the dynamic of Services Combination.

Semantic Web Services registration center is the backer of service discovery.During service registry information on services is write in the registration table, meanwhile utilize OWL-S that semantic tagger is carried out in Web service,, improve the recall ratio and the precision ratio of service to support semantic query operations such as refinement, extensive, sane level expansion.The service of registration can be a simple services, also can be the composite services that encapsulated interdepartmental complex business process.

Be based on the formalized description of XML by the abstract operation flow of flow process modeling generation.Utilize mating of procedure information in OWL-S packing service demand and the operation flow database, comprise process name coupling, data-interface coupling and constraint condition coupling.As there not being the flow template of demand, the user can manually set up and it is imported the flow process storehouse.

The service semantics coupling is the basis of service discovery, in the Services Combination process of man-machine coordination, the semantic matches of service requires more flexible, and the user can be according to actual conditions and the preference custom fit of oneself, as setting the weights of coupling satisfaction, occurrence and each occurrence.

The composite services engine is in process of analysis, composite services process, and each the discovery is integrated into operation flow with it after the satisfied service.The search strategy of user customizable service comprises that local optimum is searched with global optimum to search.Local optimum is searched and is meant from the process starting point or two ends, all seeks the service of mating the most with existing procedure at every turn.Global optimum searches, and promptly according to abstract service logic, finds one or the suitable service of a combination for each node, calculates the degree of association of serving between adjacent node, adopts dynamic programming algorithm to find out globally optimal solution at last in the state space that generates.

Adopt OWL-S as the semantic Web service combination framework, utilizing domain body to provide semantic supports, realize the automatic binding of precursor service and follow-up service interface in the automatic binding, flow process of abstract flow process and active service, and data conversion between the interface and transmission, as the conversion of data dimension.Finish the automatic monitoring and the execution of composite services at last, and the service data of flow process is gathered storage in real time.

(3) advantage and effect:

1, the present invention proposes under semantic Web service combination framework, serve intellectuality by introducing semantic making, and then finish the establishment of combination process in the mode of man-machine coordination based on abstract operation flow.This method efficiently solves the uncertainty of existence in the current AI planning and the problem of complexity problem and quiescent operation stream bad adaptability, thereby improves the success ratio and the efficient of Services Combination.

2, the present invention has provided and has utilized body (OWL) and Chinese words segmentation to carry out the intelligent search of Web service and the method for coupling, efficiently solves the restriction of the imperfection of domain body to the service semantics coupling.

3, the present invention has provided the Web service registration center that supports semantic query, the user can carry out accurate description to the target of required inquiry, can use semantic operations such as refinement, extensive, sane level expansion to inquire about and can measure matching degree between candidate and query aim, the recall ratio of service and precision ratio all can improve a lot.

4, the present invention has provided the interactive services semantic matching method, be that man-machine Shared Decision Making is selected service, make service search more accurate, also easier finding had complementary functions and the basic service of Interface Matching, thereby obtain the matching result of fancy grade, improve the efficient of Services Combination.

5, the present invention has finished the graphical modeling of operation flow and designing and developing of Services Combination engine, and utilizes FLEX/AS/JSP/JAVA to develop interactive services combined system based on abstract service logic.

Description of drawings

Fig. 1 is a Services Combination flow process frame diagram;

Fig. 2 is three layers of map of services illustraton of model;

Fig. 3 is a services request customization surface chart;

Fig. 4 is that abstract operation flow is checked surface chart;

Fig. 5 is abstract operation flow figure modeling surface chart;

Fig. 6 is a Semantic Web Services coupling binding surface chart;

Fig. 7 is that operation flow node service matching is chosen surface chart.

Embodiment

The present invention includes a kind of interactive Semantic Web Services dynamic combined method based on abstract service logic, this method mainly is under the semantic Web service combination framework, based on the abstract service logic in the manufacture course of products, abstract business description and concrete Web service example are separated, and improve the efficient of Services Combination, man-machine coordination dynamic creation operation flow in conjunction with the Semantic Web Services technology.Fig. 1 has provided Services Combination flow process framework, and included gordian technique has operation flow modeling, Semantic Web Services registration center, operation flow coupling, service semantics coupling, service search strategy and service dynamic binding.

As shown in Figure 2, Services Combination framework of the present invention is from being divided into three layers in logic.The abstract Business Logic of the superiors utilizes on the flow process modeling tool macroscopic view and has made up the Services Combination structure, has described sequential relationship and control relation between each node.This level does not need to consider whether this flow process relates to different departments different platform, how effectively the organizational resources service of only paying close attention to form powerful collaborative design flow process, and the change of flow process and concrete service realize irrelevantly, thereby can realize that the flexible customization of operation flow satisfies users ' individualized requirement.Be positioned at the platform that the service registry storehouse of the second layer provides a service registry and searched, utilize OWL-S that unified semantic description is carried out in Web service, shielded the difference that service realizes between the different departments different platform, the lotus root that has reduced between abstract business description and the concrete service realization is right.Described service can be a simple services, also can be the composite services that encapsulated interdepartmental complex business process.The concrete service of the bottom realizes it being final executor movable in the business process model, they with the service registry storehouse in OWL-S service describing realization static binding.Below in conjunction with accompanying drawing Services Combination process of the present invention and gordian technique thereof are described in further detail.

The first step is set up Semantic Web Services registration center, finishes the semantic tagger of service, adds service QoS information, has generated the service ontology that utilizes OWL-S to describe at last.The user carries out service registry by submitting the WSDL service description file to.The user can make amendment to service name, service describing, input and output in the registration process, also want to add the associated person information of service, such as name, unit, academic title, phone, fax and Email etc., in the service ontology storehouse, generate the service ontology file that utilizes OWL-S to describe at last, and its URL address is write web services registry.

In second step, the user formulates the services request scheme according to the preference of oneself.Services request is encapsulated the services request body that generation utilizes OWL-S to describe, and carry out semantic matches with the service in the service ontology storehouse.

At first, as shown in Figure 3, the user can be that some node is determined occurrence flexibly according to actual conditions, as Service name, and classification, input, QoS etc.The weights of each occurrence of dynamic assignment as required in addition.When having a plurality of intimate candidate service, need the filtration of serving by the qos value of service.QoS is devoted to serve synthetic operability rather than the synthetic function of service.The weights of the same every index of QoS also can dynamically be formulated.

Secondly, if can not find the service of meeting consumers' demand in the service library, then user's request and the combination process in the operation flow storehouse are done coupling.Sort by matching degree and provide the abstract operation flow of system recommendation in abstract business description in the operation flow storehouse and services request object Profile coupling back, the user then selects corresponding flow process to carry out Services Combination.

In the 3rd step, the user clicks corresponding flow process and can check the graphical demonstration of flow process and the descriptor of flow process as shown in Figure 4, comprises process name, affiliated field, creates unit and creation-time etc.Click the descriptor that the node in the process flow diagram also can be checked node, if this flow process is satisfied the demand then started Services Combination.

Can't satisfy actual needs as current flow process, then can utilize the flow process modeling tool that this flow process is edited.As shown in Figure 5, the user can edit flow process descriptor and nodal information, also can edit the port information of node, and wherein port information comprises port title and data type.In the time of the process flow diagram editor, the XML source code of system's product process.

In the 4th step, the semantic matches of service is the basis of service discovery and combination, and it then is the gordian technique of service semantics coupling that the concept matching degree calculates.

At first, utilize domain body to set up the notion layering, and adopt the thought of semantic distance to calculate matching degree between the notion, the computing formula of notion semantic matches degree is as follows:

Wherein, c ₁And c ₂Represent source notion and purpose notion respectively, level represents the residing level of notion, and depth represents the degree of depth of conceptional tree, i.e. the maximal value of all node levels.

The calculating of notion semantic matches degree has directivity among the present invention.If source notion c ₁Be purpose notion c ₂Parent or similar, c then is described ₂Must be to satisfy c ₁, its matching degree is 1; If source notion c ₁Be purpose notion c ₂Subclass, c then is described ₂The notion covering scope is greater than c ₁, 1-(int (c ₁)-int (c ₂)) then represented c ₂Satisfy c ₁The size of possibility; If source notion c ₁With purpose notion c ₂There are not equivalence or inheritance, c then is described ₂Necessarily can not satisfy c ₁, thereby its matching degree is 0.

Secondly, because the imperfection of domain body is long-standing, so have hysteresis quality aspect the representation of knowledge reasoning.In actual central continuous expansion along with quantity of service, new knowledge constantly occurs, and the user has randomness aspect the information on services expression.The present invention has introduced Chinese words segmentation on the basis of calculating based on the notion semantic matches of body, calculate the concept matching degree under the support of specialized dictionary.Utilize Chinese words segmentation to cross to filter the function word in the notion, the specialized vocabulary in the middle of extracting obtains source notion vector S respectively _mWith purpose notion vector D _n, calculate S afterwards _mAnd D _nIn semantic matches degree between each notion obtain mating matrix M D _{M * n}, draw the matching degree computing formula at last:

$\mathrm{md}(s,d)=\left\{\begin{array}{cc}\frac{m}{n}\×\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{1\≤j\≤n}{\max }{\left(\mathrm{MD}\right)}_{\mathrm{ij}},& n>m\\ \frac{n}{m}\×\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{i\≤j\≤n}{\max }{\left(\mathrm{MD}\right)}_{\mathrm{ij}},& n\≤m\end{array}\right.$

$=\left\{\begin{array}{cc}\frac{m}{n}\×\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{1\≤j\≤n}{\max }\left[\mathrm{sim}(S_i,D_j)\right],& n>m\\ \frac{n}{m}\×\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{1\≤j\≤n}{\max }\left[\mathrm{sim}(S_i,D_j)\right],& n\≤m\end{array}\right.$

Wherein, m, n represent source notion vector S respectively _mWith purpose notion vector D _nDimension.

At last, set up the model of service describing: [base attribute, input parameter, output parameter, QoS], wherein Fu Wu base attribute mainly is meant service name, classification of service and service registry person.Wherein classification of service and serve the registrant and be used for service to be matched is screened overanxious does not participate in the calculating of service semantics matching degree.Therefore on the basis of calculating the concept matching degree, can obtain the computing formula of service semantics matching degree:

MatchingDegree(WSR，WS)＝ω ₁×matching(WSR _name，WS _name)+ω ₂×matching(WSR _in，WS _in)+ω ₃×matching(WSR _out，WS _out)+ω ₄×matching(WSR _qos，WS _qos)

ω wherein ₁+ ω ₂+ ω ₃+ ω ₄=1, WSR represents services request, and WS represents service to be matched.

The coupling of service name can be tried to achieve by calculating the concept matching degree; It is the maximum match vector M atchVector that cum rights bipartite graph maximum match obtains the parameter interface that the coupling of input and output parameter is then utilized the KM algorithm, has considered the difference that the number of parameters difference causes simultaneously, draws the matching degree computing formula of input and output parameter at last:

$\mathrm{matching}({\mathrm{WSR}}_{\mathrm{ParaList}},{\mathrm{WS}}_{\mathrm{ParaList}})=\frac{({\left|\right|\mathrm{MatchVector}\left|\right|}_1/n+{\left|\right|\mathrm{MatchVector}\left|\right|}_1/m)}{2}---\left(5\right)$

M wherein, n represents source parameter and target parameter dimension respectively, || MatchVector|| ₁1 norm of expression maximum match vector, i.e. the absolute value sum of each value in the vector.

QoS descriptive model among the present invention comprises initial time delay, carries out time delay, expends, availability and reliability five indices, draws the computing formula of QoS information matches thus:

matching(WSR _qos，WS _qos)＝ω ₁₁×matching(WSR _it，WS _it)+ω ₁₂×matching(WSR _pt，WS _pt)+ω ₁₃×matching(WSR _cost，WS _cost)+ω ₁₄×matching(WSR _avl，WS _avl)+ω ₁₅×matching(WSR _rel，WS _rel)

ω wherein ₁₁+ ω ₁₂+ ω ₁₃+ ω ₁₄+ ω ₁₅=1.

The coupling of qos parameter is divided into cost shape parameter coupling and benefit shape parameter coupling, and expending (cost) as the cost shape parameter with service is that its matching formula of example is:

$\mathrm{matching}({\mathrm{WSR}}_{\cos t},{\mathrm{WS}}_{\cos t})=\left\{\begin{array}{cc}1,& {\mathrm{WSR}}_{\cos t}\&GreaterEqual;{\mathrm{WS}}_{\cos t}\\ {\mathrm{WSR}}_{\cos t}/{\mathrm{WS}}_{\cos t},& {\mathrm{WSR}}_{\cos t}<{\mathrm{WS}}_{\cos t}\end{array}\right.$

The benefit shape parameter is that its matching formula of example is with availability (availability):

$\mathrm{matching}({\mathrm{WSR}}_{\mathrm{avl}},{\mathrm{WS}}_{\mathrm{avl}})=\left\{\begin{array}{cc}1,& {\mathrm{WSR}}_{\mathrm{avl}}<={\mathrm{WS}}_{\mathrm{avl}}\\ {\mathrm{WS}}_{\mathrm{avl}}/{\mathrm{WSR}}_{\mathrm{avl}},& {\mathrm{WSR}}_{\mathrm{avl}}>{\mathrm{WS}}_{\mathrm{avl}}\end{array}\right\}$

The 4th step, after determining, flow template then carries out the service semantics coupling, and give one or one group service of each node binding as shown in Figure 6, and calculate the matching degree of interface between the adjacent node corresponding with service.

Click selected service and can check the detailed description information of service, comprise information such as title, description, classification, input, output, contact person and QoS.Except the static information that can check service, the user also can call performance and the function of selected service with Analysis Service.

Do not satisfy the demand as the service of binding on the node, the user can adjust according to the preference of oneself, as shown in Figure 7, can edit the occurrence content, and it is overanxious to increase QoS, and the priority that each occurrence is set respectively as required calculates weights, and shows with cake chart.After service matching was finished, the service of clicking in the service list can utilize histogram to check the matching degree of each occurrence and the matching degree after the weighting.

In the 5th step, carry out combination process.

At first, describe the graph structure that draws flow process, draw the sequential relationship and the control relation of each node of flow process again according to the adjacency matrix of figure by resolving abstract operation flow.

Secondly, be that the basis utilizes the KM algorithm to obtain maximum match between parameter vector with the notion semantic matches, thus the automatic binding of precursor service and follow-up service interface in the realization flow.

At last, set up the experiment process database, and the service data of flow process is gathered in real time and stored.Under the support of dimension body, finish data conversion and transmission between the interface in the process of implementation, as the conversion of data dimension.The node data that the user can be provided with in the monitoring rank convection current journey is monitored.

Claims (17)

1. interactive Semantic Web Services dynamic combined method based on abstract service logic, this method may further comprise the steps:

Step 1) is resolved the services request scheme that the user formulates, and makes up the request body, and mates with the service of service register center, if the service that does not satisfy the demands then the combination process storehouse is mated automatically in system provides feasible combination process template;

Step 2) if the resulting flow template of step 1 is not met consumers' demand, then the user edits described flow template or newly-built as required combination process by graphic interface;

The step 3) system carries out service search and coupling according to the descriptor of flow process, and perhaps the user adjusts the service matching of respective nodes in good time according to the preference of oneself;

Step 4) will be served with abstract operation flow according to the matching result of step 3 and be carried out dynamic binding, generate combination process;

The combination process that step 5) execution in step 4 generates is monitored the flow nodes implementation status, and the data that operation produces are gathered storage in real time.

2. according to the said method of claim 1, it is characterized in that: said services request scheme comprises service name, classification of service, registrant, input parameter, output parameter and service quality (QoS) information in the step 1.

3. according to the said method of claim 1, it is characterized in that: said request body is to utilize the OWL-S based on body that services request is encapsulated in the step 1, and its Profile object and registration service are mated.

4. according to the said method of claim 1, it is characterized in that: the service register center in the step 1 comprises service essential information concordance list, registered user's table and service sorted table.

5. according to the said method of claim 1, it is characterized in that: combination process in the step 1 coupling is to utilize Hibernate to carry out the persistence encapsulation flow database and mate calculating with the services request object.

6. according to the said method of claim 2, it is characterized in that: the quality of service information in the step 1 in the services request scheme comprises initial time delay, carries out time delay, expends, availability and reliability.

7. according to the said method of claim 3, it is characterized in that: in the step 1 the Profile object factory profile information of service, when mating, only need obtain the URL address of service to be matched with registration service, calculate matching degree then with services request.

8. according to the said method of claim 4, it is characterized in that: the service essential information concordance list of service register center comprises service name, service describing, classification of service, service registry people, service registry time and service URL address in the step 1.Classification of service and serve the registrant and when service matching, be used to treat match service and screen.

9. according to the said method of claim 5, it is characterized in that: the abstract combination process in the step 1 and the coupling of services request object are calculated and are comprised process name coupling, data-interface coupling and constraint condition coupling.

10. according to the said method of claim 1, it is characterized in that: for described flow template is edited or newly-built as required combination process is by patterned modeling tool of Flex technological development in the step 2, draw out node and sequential relationship and control relation in the flow process by manual mode, and generate the understandable formalized description of computing machine.

11. according to the said method of claim 10, it is characterized in that: the essential information that the flow nodes in the step 2 need be added or edit comprises nodename, node type, affiliated field, node description, input port and output port.

12. according to the said method of claim 10, it is characterized in that: the procedural model formalized description is meant the XML description document of product process self after the flow process correctness obtains checking in the step 2.

13. according to the said method of claim 1, it is characterized in that: the flow process descriptor comprises title, issuing unit, issuing time, affiliated unit, input and output in the step 3; Step 3 further is included as some nodes and determines occurrence, constraint condition flexibly, and the weights of each occurrence of dynamic assignment.

14. according to the said method of claim 1, it is characterized in that: binding is that abstract operation flow is resolved the logical topology structure that draws flow process described in the step 4, each node is all bound one or one group of service, calculate the degree of association of serving between adjacent node respectively, asking for optimal path in the state space that generates, serves as that the basis connects each atomic service with described optimal path; Combine engine is bound service automatically, generates to satisfy the required all or part of combination process of user, and the user adjusts repeatedly according to the preference of oneself on this basis, and each node all finds suitable service in flow process.

15. according to the said method of claim 14, it is characterized in that: it is to obtain the maximum match of precursor service and follow-up service according to the KM algorithm that the service in the step 4 connects, and then finishes the automatic butt of service data interface.

16. according to the said method of claim 1, it is characterized in that: the implementation of the combination process in the step 5 can be monitored in real time, the user is provided with the monitoring rank varigrained service is monitored.

17. according to the said method of claim 16, it is characterized in that: the varigrained service of monitoring is the composite services that comprised in the flow process in the step 5, only monitors outer service data or monitor its inner institute to comprise atomic service.

Images