Intelligent Outage Probability Prediction for Mobile IoT Networks Based on an IGWO-Elman Neural Network

Advances in sensor technology have accelerated the development of mobile internet applications and contributed to the tremendous growth of the Internet of Things (IoT). Mobile IoT networks typically operate in complex and highly variable urban wireless channel environments. This makes effective and reliable communications challenging. To ensure efficient and stable communications, mobile IoT networks must adapt to the complex environment. This can be achieved by predicting the outage probability (OP). This paper investigates OP analysis and prediction for these networks. Exact closed-form OP expressions are derived and Monte-Carlo simulation is used to verify the analysis and evaluate the OP performance. Then, an OP prediction approach based on an improved grey wolf optimization (IGWO) algorithm and Elman neural network (IGWO-Elman) is proposed. The IGWO algorithm employs improved opposition-based learning to optimize the population initialization and ensure sufficient population diversity. The Elman neural network uses the IGWO algorithm to obtain good network parameters. Simulation results are presented which demonstrate that the prediction accuracy of IGWO-Elman is better than SVM, ELM, and BP algorithms. In terms of prediction accuracy, the IGWO-Elman algorithm is increased by 44%.

Automated summary

Advancements in sensor technology have accelerated the development of mobile internet applications and contributed to the significant growth of the Internet of Things (IoT). This paper focuses on outage probability (OP) analysis and prediction for mobile IoT networks, proposing an OP prediction approach based on an improved grey wolf optimization (IGWO) algorithm and Elman neural network. Simulation results show that the prediction accuracy of IGWO-Elman outperforms SVM, ELM, and BP algorithms, with an increase of 44% in prediction accuracy.