Ontology Semantic Web Services Composition Based On Al Planning

With the rapid development of internet techniques, Web services are becoming more and more widely used. Web services are a technology that enables provision of functionality, on the application level, or on the business level, by means of a standardized interface in a way that they are easily invoked via internet-protocols. As the capability of one individual Web service is limited, it's necessary to create new functionalities that can automatically compose existing semantic-based Web services together to fulfill the practical requirements.Composition of Web services has received much interest. On the one hand, many organizations have developed a number of XML-based standards to formalize the specification of Web services, although this approach is primarily syntactical. On the other hand, the Semantic Web community focuses on reasoning about web resources by explicitly declaring their preconditions and effects with terms precisely defined in ontologies. So far, both approaches have been developed rather independently from each other. We want to find a new solution to compose Web services using AI planning.To meet this goal, we have to resolve two problems: First, service is usually developed by different organizations, they use different conceptual models to express the various features of services. In order to automatically compose Web services, we have to require a unified description of service model to express their relevant semantic information. The Semantic Web provides a processing level description of the service which models the preconditions and post conditions of the process so that the evolution of the domain can be logically inferred. It relies on ontologies to formalize domain concepts which are shared among services. Second, the service environment is a highly flexible and dynamic environment. New services become available on a daily basis and the number of service providers is constantly growing. Ideally, service processes should be able to transparently adapte to environment changes, and to adapt to customer requirements with minimal user intervention. A larger trend is using artificial intelligence planning to deal with automatic portfolio, through the rules of reasoning and semantic matching to create a job stream.In this paper, we use AI planning strategies to solve the problem of Web Services composition. We define translation templates PXML to translate OWL-S to PDDL; discuss the method of semantic Web services composition based on AI planning; implement a Web Service composition system PTRBM.Translation templates PXML can convert the domain ontology and service descriptions in OWL and OWL-S, respectively, to equivalent PDDL problem and domain descriptions. The hasPrecondition service parameter can be transformed to the precondition of the action by use of predicates directly. The same holds for the hasEffect condition parameter. Mapping the service input parameter X to a special type of the service precondition parameter. Each input variable X is mapped to an input parameter X of an PDDL action complemented by precondition predicate"agentHasKnowledgeAbout(X)". The same with the output parameters. We defined associated node, which tells the relation between the nodes, the link construct is used to express these synchronization dependencies. It is necessary to have some preconditions, like whether the required data has been formed, what the status should be. And after the associated node has been done what is the effect and what we should do.Method of semantic Web services composition based on AI planning: First of all, preprocess the input data assigning initial values to each predicate of the given state descripted in PDDL. We create a tree structure, which describes all type relations and type inclusions. Predicates which are neither negative nor positive in the effect list of an operator are removed from all preconditions and effect lists. So, the preconditions and effects are converted to disjunctive normal form. After the domain and problem definitions have been parsed, we find that if we can produce a plan sequence of actions that solves a given problem in a simple form of hierarchical task networks. When there is a plan, concrete set of methods and decomposition rules that lead to an executable plan is provided. If we would not find the solution due to the fixed structure of hierarchical action decompositions stored in its database, we should use a similar method like graph-plan with a function of calling back. This method constantly analyze according to the current status which action can be implemented and record the effect of the change after the implementation. Each layer will be expanded from the initial state using an enforced hill-climbing search algorithm to determine whether there is a new action can not become satisfied because of the implementation of action. Repeat it until we find a solution or there is nothing can change while using the action that we can use. If there is a plan, return the result; or return can not find a plan.Finally, a Web services composition system PTRBM is implemented, its main capabilities include: (1) shows the original composition plan generated by the PTRBM upon the existing problem description and initial description. (2) You can create a new type by selecting the type of which the newly created type is subtype. (3) You can change some predicate definitions in the planning domain. (4) You can change the definition of any of the OWL-S services that is associated with the actual planning domain and problem. (5) You can modify your loaded problem description. (6) You can do a post optimization of the composed plan with respect to a set of given quality of service. Explain how to composite the benefits of online shopping account services and analysis of the relevant background; discuss the relationship between the required object types. Using the algorithm provided in this paper to generate a graphical results output. PXML translation template will translate the OWL-S language which lack the ability to describe the rules to PDDL language, using AI planning will basically meet the requirements of Semantic Web Service Automatic composition. But in the method provided in this paper, in the module of data preprocessing, we only think about the relation of inheritance. In this way, it is convenient to find an object to replace the one we need. If two objects extend form the same father object, then there is possibility that under certain circumstances they are replaceable. If there is only combination relationship, no inheritance relationship between the services, they should not substitute for one another. In the future, the way to find the alternative service will be improved, using the concept of the degree of similarity to determine whether the service can be used.