Research On QoS-driven Service Composition Optimization Algorithm In Internet Of Things

With the development of the technique of Internet of things,more and more intelligent embedded devices have blurred the boundaries between the virtual world and the real world in commercial applications.Most of the intelligent devices in the Internet of things environment have communication and computing power,which can provides Web Service based Internet of things services.Recent years,with the increase of the number of Web services based Internet of things services,the number of Web services with the same function is increasing,but the Web services with the same function usually have different QoS attributes,which leads to the expansion of the Web service composition and the increase of the difficulty.Therefore,how to efficiently combine individual atomic services into a composite services with high QoS becomeing the research hotspot of Web service composition.Because the genetic algorithm can effectively solve the problem of large-scale optimization,which has been widely used in the optimization of service composition.But the fixed cross probability and mutation probability of traditional genetic algorithm is easy to result to the premature convergence problem.Therefore,this thesis improves the traditional genetic algorithm by introducing the activation function in the neural network and a variable which can comprehensively describe the diversity of the population and the evolutionary algebra,then proposed an imporved genetic algorithm which can adaptively adjusts the cross probability and the mutation probability according to the diversity of the population and the number of iterations.Compared with the traditional genetic algorithm,this algorithm is more consistent with the natural law of population evolution and has the ability to jump out of the local optimal solution.In solving the large-scale service composition optimization problem,it can effectively avoid premature convergence and produce higher fitness individuals.The simulation experiments show that the improved genetic algorithm has a strong global optimization ability compared with the traditional genetic algorithm,and is more suitable for solving the large-scale Internet of things service combination problem.