A DQN-based model for intelligent network selection in heterogeneous wireless systems

TL;DR

This paper presents a DQN-based reinforcement learning model for intelligent network selection in heterogeneous wireless systems, improving accuracy in choosing optimal RATs over traditional methods.

Contribution

It introduces a novel DQN-based approach for dynamic network selection, enhancing decision accuracy in heterogeneous wireless environments.

Findings

Achieves 93% accuracy in network selection

Outperforms traditional MADM methods with up to 75% accuracy

Effective in switching between multiple wireless technologies

Abstract

Wireless communications have been at the center of the revolution in technology for the last few years. The 5G communication system is the pinnacle of these technologies; however 4G LTE, WiFi, and even satellite technologies are still employed worldwide. So, the aim of the next generation network is to take advantage of these technologies for the better of the end users. Our research analyzes this subject and reveals a new and intelligent method that allows users to select the suitable RAT at each time and, therefore, to switch to another RAT if necessary. The Deep Q Network DQN algorithm was utilized, which is a reinforcement learning algorithm that determines judgments based on antecedent actions (rewards and punishments). The approach exhibits a high accuracy, reaching 93 percent, especially after a given number of epochs (the exploration phase), compared to typical MADM methods where the accuracy does not exceed 75 percent