Abstract

One of the advantages of smart phones is the ubiquity of sensing and computing power. The smart phone consists of a mobile computing platform. The Web browsers installed on the platform make it possible to surf the Internet through mobile broadband and Wi-Fi. An important aspect of smart phones is that they have application programming interfaces, which is able to take advantage of third-party applications. Different from any desktop applications, the smart phone applications could be highly adaptive to contexts, i.e. according to context information, e.g. location, identity, and time, the applications are tuned to satisfy particular requirements in the contexts. On the other sense, service composition is a way to plan a business process to fulfill business goals that cannot be achieved by individual business services. Service composition can be modeled as a AI planning problem. Based on the initial context and the goal context, planning-based service composition launches a goal-oriented composition procedure to generate a plan. Service composition over smart phones can be context-awareness. In this thesis, we want to investigate context based service discovery and service composition over smart phones. We propose a constraint-based context model. We include non-electronic services into service composition, which extends the scope of services considered in existing service composition research. Moreover, our composition algorithm suits mobile computation power because the service composition can adjust to the computation power of mobile phones easily. As
a motivating example, we build an entertainment planner over an Android phone.