A Value-Driven Approach for the Determination of Global Optimization Objective of Service Composition

We present a value-driven approach capable of determining the global optimization objective of service composition to satisfy high-level value expectations. In traditional service composition methods, global and local QoS requirements and constraints are directly raised by customers, then a set of service components (web services) are selected to achieve the optimized QoS. Yet in practice, there is no mechanism helping customers accurately and comprehensively determine their QoS desires. In our approach, a customer´s initial expectations of high-level global values are first put forward, and then five types of dependencies between service values are analyzed. Because values are realized under the facilitation of specific service components, value expectations and value dependencies are transformed into the objective and constraints of service selection and composition, respectively. Another feature of our approach is that, besides web services, other generalized service components such as human activity and physical resources are considered as potential components for fulfilling value expectations. The effectiveness of our approach is demonstrated in a case study of ocean transportation service (OTS).