The Application Of Description Logic In Component Retrieval And Composition Of Web Services

Description logic is a formalization facility of knowledge representationbased on object, also named concept representation language orterminological logic. It is a decidable subset of first-order predicate logic. Itpossesses suitably defined semantics and strong expressive ability. Thedifference between description logic and first-order predicate logic is thatdescription logic system can provide decidable inference services. Theimportant features of description logic are that description logic possessesstrong expressive ability and decidability, assures the termination of inferencealgorithm and returns the right result. Satisfiability is the kernel problem ofdescription logic inference.Among the numerous formalization methods of knowledgerepresentation, description logic gets especial attention in these 10 years, themain reasons are that it has clear model theory mechanism, is fit forrepresenting application domains through concept taxonomy and providesmany inference services.A description logic system contains four basic parts: construction setrepresenting concepts and roles; TBox, inclusion assertion; ABox, instanceassertion; inference mechanism on TBox and ABox.In a general way, description logic constructs complex concepts androles based on simple concepts and roles according to the providedconstruction arithmetic operators. Generally, description logic contains thefollowing arithmetic operators: intersection (?), union (?), negation (?),existential quantification (?) and value restriction (?) at least. This kind ofthe most elementary description logic is called ALC. Adding differentconstruction arithmetic operators to ALC can form description logic withdifferent expression ability. Description logic is used as knowledge representation facility in manydomains. With the in-depth study of description logic in these 10 years,description logic has been applied to semantic web, software engineering,configuration, digital library, database, concept modeling, query optimizationand view maintenance, natural language semantic, intelligent informationintegration, information access and intelligent interface, formalizationcriterion of project, terminology and ontology, planning, etc.Based on the in-depth study of basic theory, non-standard inferences andexisting applications of description logic, the paper mainly researches on theapplications of description logic to component retrieval and web servicecomposition, and implements the correlative algorithms.Reuse is a process of matching between reuse requirement and productdescription. The description and retrieval of component is the emphasis of theresearch on software reuse and component library. Different componentretrieval methods are based on the same component description scheme.Component description and retrieval based on facet is one of the mostextensively used methods. Facet classification method describes complexobject from different dimensions, consequently enhances descriptionaccuracy. The method possesses good retrieval efficiency, but it also haslimitations. The main problem of component reuse is the incompleteness ofrequirement and product description. In component retrieval method based onfacet, reuse requirement of user is incomplete and imprecise in manycircumstances. There are two reasons about this. The first reason is theambiguity of user reuse requirement. The second reason is the limitedrequirement expression means that are provided to the user. From the view ofcomponent description, facet scheme can describe a component relativelycomprehensive. But it also has some limitations. For example, the mainrelationship among terms is generic special relationship, facet scheme cannotexpress cross classification among terms. Therefore, simply using descriptionscheme based on facet cannot support user to sufficiently express retrieval requirement, the description information of component is also incomplete.Based on description logic, it is able to describe user reuse requirementmore exact and more complete, and is also able to provide rich semanticcomments to facet description of component. Consequently, effectively closethe gap between user reuse requirement and component description.On the basis of facet-based description method, the paper introducesdescription logic as formalization facility of knowledge representation todescribe user reuse requirement and component. Consequently, make themmatch on the semantic level. The paper introduces concept called matchingdegree to denote the matching extent between user reuse requirement andcertain component in the component library. The paper proposes an algorithmfor component retrieval and matching based on description logic, thealgorithm can calculate a real number called penalty to associate matchingdegree between user reuse requirement and certain component, the smallerthe penalty, the better the matching degree. Consequently, help user selectsuitable component from the component library. In the algorithm, because ofthe help of domain ontology, it is possible to judge the relationship amongconcepts when retrieving components. Compared to traditional componentretrieval method, the recall rate and precision rate of component retrievalincrease. The algorithm can sort the result of component retrieval accordingto exact match, full match and potential match. In the same match type, theretrieved component can be sorted according to the value of penalty.Consequently, provide the user a basis for choice.With the increasingly mature of web service technology, more and morestable and easy-to-use web services are shared on the Internet. But single webservice can only provide limited function. In order to more fully utilize theshared web services, it is necessary to compose the shared web services toprovide stronger service function, quicken the speed of system developingand fast satisfy user requirement. Web service composition provides to usernew and value-added service by composing simple web services. That is, compose existing web services to form web service which possesses strongerfunction and can solve complex problem. When the required functions in realcircumstances become more and more complex and single web servicecannot satisfy user requirement, it is necessary to use web servicecomposition technology to form new service and provide more complexfunction. At present, people have already done a lot of work on web servicecomposition.The description of service using existing protocol and standard is limitedto syntax level, cannot express semantic information. Therefore, current webservice composition is executed by manual method. Manual method istime-consuming and effort-consuming, and is difficult to fit the dynamicfeature of service. So it is necessary to research on how to implementdynamic composition of web services based on semantic.Description logic can accurately describe concepts and relationshipsamong concepts. Therefore, it becomes the semantic basis of mutualunderstanding between people and machine, as well as machine and machine.On the problem of web service composition, the paper mainly researcheson how to use description logic to implement dynamic composition of webservices. The paper presents the definitions of matching degree of outputdecomposition, service relationship map and etc, proposes an algorithm fordynamic composition of web services based on description logic. Thealgorithm considers not only the semantics of services but also the qualityand efficiency of service composition. It can implement dynamiccomposition of web services according to service request of userautomatically. In the algorithm, because of the help of domain ontology, it ispossible to judge the relationship among concepts when composing webservices, so that the probability of successful composition of web servicesincreases. Compared to traditional method of web service composition, thealgorithm considers the semantic of web service. Consequently, increase thequality and probability of success of web service composition.