A weighted approach to partial matching for mobile reasoning

Abstract

Due to significant improvements in the capabilities of small devices such as PDAs and smart phones, these devices can not only consume but also provide Web Services. The dynamic nature of mobile environment means that users need accurate and fast approaches for service discovery. In order achieve high accuracy semantic languages can be used in conjunction with logic reasoners. Since powerful broker nodes are not always available (due to lack of long range connectivity), create a bottleneck (since mobile devices are all trying to access the same server) and single point of failure (in the case that a central server fails), on-board mobile reasoning must be supported. However, reasoners are notoriously resource intensive and do not scale to small devices. Therefore, in this paper we provide an efficient mobile reasoner which relaxes the current strict and complete matching approaches to support anytime reasoning. Our approach matches the most important request conditions (deemed by the user) first and provides a degree of match and confidence result to the user. We provide a prototype implementation and performance evaluation of our work. © Springer-Verlag Berlin Heidelberg 2009.