Trustworthy Metrics Method Of Semantic Web Services Composition

The rapid development of computer and network makes network application transform from computer-center to user-center, and the network applications change from“sell the software to users”to“customize trustworthy services for users”. This development trend makes a requirement that the software development would be able to discover the programs which meet user needs according to the specific requirements of different applications and do a composition of services to provide the trustworthy service for users. The proposing of semantic Web services makes users able to provide their own software to other users as a service as well as call software services of other users; relatedly, the networked software metrics helps to ensure the trustworthiness of the selected service, which ensures that the composition services can meet not only the functional users needs but also performance users needs. It can be seen that user demand-drive, trsutworthy assurance service composition in the networked environment has become one of key research focus in the services computing and trusted computing.However the existing method of service composition is not enough for the trustworthiness, especially the related work in the service composition process is very few. The weakness of such studies affect the extensive application of semantic web services, and more satisfactory solution to this problem would be a new and interesting research topic.In this paper, some knowledge and methods such as domain ontology, preferences, fuzzy description logic, semantic similarity and defect metrics are used to describe and solve the trustworthy metrics of semantic web services composition, and the main work and achievements are as follows:?to analyze the trustworthy properties of services compositionAimed at the existing problem of the blurred trustworthy properties of services composition, Composition Services Tree (CST) is designed to clear the trustworthy properties to divide the basic form of services in the composition into atomic services and composition services: the trustworthy properties of atomic services are relatively static, so they can be designed by reference to the trustworthy properties of general software; the trustworthy properties of services composition are more dynamic, so they are parsed by using Trustworthiness Analysis of Services Composition based on Callings (TASCC). ?to build the preference ontology of services compositionAimed at the problem that services composition are hard to model semanticly, preference ontology is used to describe the semantics of services composition and provde logical support for the extraction of trustworthy metrics indicators. According to the basic framework of domain ontology, Concept Layered Model (CLM) and Multiple Ontology Learning Model (MOLM) are designed to build domain ontology to describe general semantic framework of services composition, as well as Function Preference Model based on Concept (FPMC) and Interest Preference Model (IPM) are designed to add preference attributes into built domain ontology in order to build the preference ontology to describe personalized semantic elements of services composition.?to establish the trustworthy metrics index system of services compositionAimed at the problem that trustworthy metrics indexes are hard to extract directly and indexes values are difficult to calculate positively, Two-dimensional Trustworthy Metrics Indexes System (TTMIS) is designed to deivide the trustworthy metrics indexes into two stages of design and implementation. Hierarchical Static Trustworthy Metrics Index System for Services Composition (HDIS) and Trustworthy Metrics Process Index System by Hybrid Modeling (TMPIS-HM) are designed to extract the trustworthy metrics indexes of services composition, as well as user satisfaction factor is added to take full account of the special influence of user to services composition.?to propose the trustworthy metrics method of services compositionPreference ontology and trustworthy metrics indexes of services composition, Preference Ontology Oriented Trustworthy Metrics Methods for Services Composition (POTMM-SC) is proposed by adding a Trustworthiness Metric Center (TMC) which works independently in service registry to achieve scientific and objective trustworthy metrics. In POTMM-SC, indexes weights algorithms and Fuzzy Description Language for Trustworthy Degree of Atomic Service (FDL-TDAS) are desgined to calculate the basic trustworthy degree of atomic servces, and and then the static trustworthy degree of services composition is obtained by calculating the semantic similarity between target ontology and service ontology, and finallly the dynamic trustworthy degree of services composition is obtained by Local- Global Integrated Trustworthy Metrics Model (LGDTMM) to form the integrated trustworthy degree of services composition.