Knowledge-based web services for context adaptation.

The need for higher value, reliable online services to promote new Internet-based business models is a requirement facing many technologists and business leaders. This need coupled with the trend towards greater mobility of networked devices and consumers creates significant challenges for current and future systems developers. The proliferation of mobile devices and the variability of their capabilities present an overwhelming number of options to systems designers and engineers who are tasked with the development of next generation context adaptive software services. Given the dynamic nature of this environment, implementing solutions for the current set of devices in the held makes an assumption that this deployment situation is somehow fixed this assumption does little to support the future and longer term needs within the marketplace. To add to the complexity, the timeframes necessary to develop robust and adaptive online software services can be long by comparison, so that the development projects and their resources are often behind on platform support before the first release is launched to the public. New approaches and methodologies for engineering dynamic and adaptive online services will be necessary and, as will be shown, are in fact mandated by the regulation imposed by service level guarantees. These new techniques and technology are commercially useless unless they can be used in engineering practice. New context adaptation processes and architectures must be capable of performing under strict service level agreements those that will undoubtedly govern future business relationships between online parties. This programme of engineering study and research investigates several key issues found in the emerging area of context adaptation services for online mobile networks. As a series of engineering investigations, the work described here involves a wider array of technical activity than found in traditional doctoral work and this is reflected throughout the dissertation. First, a clear definition of industrial motivation is stated to provide the engineering foundation. Next, the programme focuses on the nature of contextual adaptation through product development projects. The development process within these projects results in several issues with the commercial feasibility of the technology. From this point, the programme of study then progresses through the lifecycle of the engineering process, investigating at each stage the critical engineering challenges. Further analysis of the problems and possible solutions for deploying such adaptive solutions are reviewed and experiments are undertaken in the areas of systems component and performance analysis. System-wide architectural options are then evaluated with specific interest in using knowledge-base systems as one approach to solving some of the issues in context adaptation. The central hypothesis is that due to the dynamic nature of context parameters, the concept of a mobile device knowledge base as a necessary component of an architectural solution is presented and justified through prototyping efforts. The utility of web ontologies and other "soft computing" technologies on the nature of the solution are also examined through the review of relevant work and the engineering design of the demonstration system. These technology selections are supported directly by the industrial context and mission. In the final sections, the architecture is evaluated through the demonstration of promising techniques and methods in order to confirm understanding and to evaluate the use of knowledge-bases, AI and other technologies within the scope of the project. Through the implementation of a context adaptation architecture as a business process workflow, the impact of future trends of device reconfiguration are highlighted and discussed. To address the challenge of context adaptation in reconftgurable device architectures, an evolutionary computation approach is then presented as a means to provide an optimal baseline on which a service may execute. These last two techniques are discussed and new designs are proposed to specifically address the major issues uncovered in timely collection and evaluation of contextual parameters in a mobile service network. The programme summary and future work then brings together all the key results into a practitioner's reference guide for the creation of online context adaptive services with a greater degree of intelligence and maintainability while executing with the term of a service level agreement