A Comprehensive Review of Ad Hoc Hexagonal Lattice Networks for Android-Based Wireless Communication

This paper provides the first comprehensive review of Android-based Mobile Ad-Hoc Networks (MANETs), emphasizing platform-specific deployment challenges and integrating bio-inspired optimization with deep learning approaches. It focuses on generic routing or simulation-only studies. This work addresses Android device heterogeneity, energy constraints and real-time adaptability, offering a practical perspective for real-world implementation. Then, it compares classical routing protocols (AODV, DSR, and OLSR), optimization-based methods (PSO, ACO, GA, and BOA), and deep learning models (LSTM, CNN, and Capsule Networks), culminating in the CapsuLSTM-BOA framework. This hybrid approach leverages Capsule Networks for spatial context, LSTM for temporal prediction, and BOA for adaptive route optimization. Performance analysis shows CapsuLSTM-BOA achieves the lowest delay of 180[Formula: see text]ms, highest packet delivery ratio of 96%, lowest energy consumption and fastest convergence speed, outperforming traditional and AI-based methods. Finally, it concludes with future directions, including federated learning, graph neural networks, and edge deployment strategies.