A Deep Learning-Based Routing Approach for Wireless Mesh Backbone Networks

Optimal routing decisions are key in communication network environments to minimize bottlenecks such as traffic congestion and limited bandwidth. Routing wireless mesh backbone networks is the focus of this study given that they popular particularly for providing broadband connectivity a huge number users accessing transmitting multimedia data hence susceptible communication-oriented bottlenecks. Existing solutions mainly optimistic approaches. These depend on link states, distance hop counts which present generalization bottleneck especially because it difficult get entire footprint network. Simply put, very determine an optimal route (WMN) with dynamic conditions. Since deep learning has strong ability. In paper, learning-based approach proposed goal ensuring defined quality service (QoS) WMN. order achieve purpose, simulation environment built feature set orchestrated generate used train Long short Term Memory (LSTM)-based model estimates QoS. The generated validated by training other models including Multilayer Perceptron (MLP), Logistic Regression (LR) Random Forest (RF). Our results show routes selected LSTM-based provides best packet delivery ratio (PDR) throughput. further (MLP, LR, RF) also provide better PDR throughput compared traditional Ad-hoc On-demand Distance Vector (AODV) protocol.