Research On Intelligent Evolution For Robustness Optimization Of Internet Of Things

The Internet of Things(IoT)has been extensively deployed in smart cities,such as smart healthcare,intelligent transportation,smart oceans and smart shopping malls.With the scale of IoT increasing,some nodes fail due to inefficient energy consumption or cyber-attacks,which severely affects the communication quality of the network.The robustness of IoT is an important property to measure the stability of network topology.Therefore,constructing a highly robust IoT topology is an urgent research topic.The current robust optimization strategies of IoT topology utilize the optimization algorithms to promote the robustness of network topology for the nodes in the monitoring area,which has high robustness and can still maintain good quality of communication faced attacks.The robust optimization algorithm mostly enhanced the robustness of the scale-free network of IoT.However,the scale-free network model has good communication capability under random attacks and is vulnerable to malicious attacks.Therefore,to improve the network topology against malicious attacks of the scale-free network model,both evolutionary algorithms and genetic algorithms are used to optimize the topology robustness against malicious attacks.But the optimization algorithms require high computational overhead and have inefficient performance of optimization.To address the problem,the paper proposes a robust optimization algorithm using neural network learning model to improve the robustness of the network topology without unchanging of the degree distribution.In order to measure the robustness of the IoT topology,we use the maximum connectivity subgraph.At the same time,the data set of the network topology is preprocessed and suitable for the neural network learning model.Compared with the previous genetic algorithm,the neural network learning model-based topology optimization algorithm has a small gap in resisting malicious attacks.In addition,the algorithm has a great advantage in resisting random attacks.