A semantic-driven Integer Programming Approach for QoS-aware Dynamic Service Composition

Service-oriented specifications aim at effectively supporting run-time service discovery, composition and brokering. QoS-aware dynamic service composition relies on algorithms that synthesize the target functional work plans, while also taking into account non-functional properties (QoS) in order to satisfy end users' requirements. In this paper, we propose a semantic-driven QoS-aware dynamic composition framework leveraging on an Integer Linear Programming technique (ILP). The proposed approach exploits the expressiveness of an ontology-based service profile model, representing structural and semantic properties of service descriptions. According to the ILP model, the service composition problem is expressed as a set of functional and non-functional constraints and an objective function built on top of a reference domain representing functional, structural and semantic properties of registered service profiles. We describe a proof-of-concept implementation of the proposed approach and discuss testing results.