Semantic Service Composition with QoS End - to - End Constraints via AND/OR Graphs

In this paper we present AND/OR graphs as a unifying framework for semantic service composition that considers users QoS constraints. The main virtues of this representation among others are its ability to express semantic inference and to deal with QoS constraints from different perspectives. In addition it correctly handles multiple inputs/outputs of services, and allows high degree of automation. Once service dependencies and QoS features are formalized as AND/OR graph, we apply a search algorithm to discover composite services that considers user QoS end - to - end preferences. The implementation of a prototype system and the experimental results support our underlying hypothesis that AND/OR graphs are not only elegant and expressive formalism for addressing QoS - aware semantic service composition, but efficient as well.