Malware threat is a major hindrance to efficient information exchange on the Internet of Things (IoT). Modelling malware propagation is one of the most imperative applications aimed at understanding mechanisms for protecting the Internet of Things (IoT) environment. Internet of Things (IoT) can be realised using agent-based modelling over complex networks. In this paper a sub-netting based scale free network in IoT (SFIoT) is proposed which is an agent based (Decision maker) based modelling and simulation using Deep Reinforcement Learning. Our scale-free network model is named based on transitions as Susceptible-Infected-Immuned-Recovered-Removed (SIIRR) that developed for large-scale complex networks. The reliability of each node using Mean Time To Failure (MTTF) is investigated. The factors considered for MTTF computations are degree of a node, node mobility rate, node transmission rate and distance between two nodes computed using Euclidean Distance. The results illustrate that the model is comparable to previous models on effects of malware propagation in terms of average energy consumption, average infections at time (t), node mobility and propagation speed
Key words: Internet of things; Agent-based modelling and simulation; Modelling malware propagation; Large-scale-free networks; Deep reinforcement learning
|
