An Energy-Efficient Mechanism Of Configurable Service Composition In The Internet Of Things

The development of Internet of Things(Io T)technology has promoted the continuous improvement of information level in industry,military,transportation and other fields.All kinds of devices have certain ability of computing.They communicate and exchange information through the Internet,so that they have the ability of perception,and smart things emerged.As a single intelligence device is often heterogeneous,their function,computing power and storage capacity are usually cannot fulfill the relatively complex requirements.The collaboration and coordination of smart things is to satisfy any demand beyond the capabilities of one device.Smart things communicate with each other through network to facilitate their communication and collaboration.At present,most applications can be supported and implemented through the multiple functions provided by smart things in the Io T environment.Service discovery and composition are long-standing research topic,and many techniques have been developed in the Web/REST service domain.Along with the emergence of service,the smart things are encapsulated as an aggregation of services with multiple functions.Therefore,the collaboration between smart things can be reduced to a service composition problem.Firstly,a three-tier IoT service framework is proposed,where the functionalities provided by smart things are encapsulated into Io T services.These IoT services are categorized into service classes according to similarity of their functionalities.By calculating the invocation possibility between service classes,a service network is built and service class chains are formed using traditional semantic Web service composition technology.In the process of construction,only the functionality of services is considered regardless of other non-functional attributes.Thereafter,recommending the appropriate service composition for service class chains.Considering the temporal and spatial constraints of IoT services inherited from smart things,energy consumption and the functionality conflicts between successive Io T services,this process can be formulated as a multi-objective optimization problem.And heuristic algorithms are adopted to solve this problem.In this article,we adopt the genetic algorithm(GA),ant colony optimization(ACO)and particle swarm optimization(PSO),and evaluate the accuracy and validity by simulation experiments.Finally,comparative experiments are carried out from three aspects including the recommended results of service composition,the minimum residual energy of the smart things and load balance of the whole network.Experimental results show that in the experimental environment of this paper,PSO can obtain a relatively optimal service composition,and its performance is better than GA and ACO by keep load balance of smart things and prolong the life cycle of the whole network.