| This dissertation addresses an automatic service composition approach for relational web services.;Although widely addressed, automatically composing a web service from existing services is still a difficult problem. This dissertation focuses on the service synthesis problem for relational services. It considers answering complex queries using a set of available relational services collectively with a two step solution. We propose a service composition framework with several algorithms to achieve service composition. We present a method to reduce the time complexity of sub-query synthesis based on condition transitive matrices. The method guarantees to find a solution if there is one.;In order to manipulate complex workflows, the SmanticObject Behavior Language is introduced to provide complex workflow support during the decomposition and composition process. We also address relational operations and how to handle each of them during our service decomposition and composition process including aggregation functions, nested queries, set operations, views, and derived relations. We also analyze those cases that our algorithms cannot find a result. |
