QOS Aware Service Composition in IoT Using Heuristic Structure and Genetic Algorithm

Abstract

In the current Internet of Things (IoT) environment, the intelligent objects develop as IoT services, some having the same functionalities but different QoS attributes. The complexity of the user's requirements to fulfill a new service (i.e., composition service) is acquired from the collection of atomic services. On the other hand, fulfilling QoS constraints in composite service becomes an NP-hard problem. The proposed approach uses a Decision Tree and Genetic Algorithm (GA) for QoS aware IoT optimal service composition. The process has used two levels of QoS attributes to achieve the optimal service composition. First-level QoS attributes are applied to the Heuristic Structure (i.e., Decision tree) to eliminate unsuitable concrete services from the abstract service that reduce search space and composition time. Finally, second-level QoS is applied to GA to find the best optimal service composition. Moreover, the scalability of the proposed system is stable and more accurate compared with the regular optimal service composition.