Dynamic link utilization empowered by reinforcement learning for adaptive storage allocation in MANET

In modern wireless networks, mobile nodes often deal with the challenge of maintaining a sufficient number of data packets due to limited storage capacity within each cluster. It adversely impacts network performance by compromising data quality during transmissions. The ensuing delays, caused by data packets awaiting storage allocation, result in reduced throughput and increased end-to-end latency. To effectively address these issues, we present a Dynamic Link Utilization with Reinforcement Learning (DLU-RL) method, which is designed to optimize storage allocation for communication data packets, significantly enhancing network performance. Instead of static allocation, DLU-RL employs dynamic strategies guided by reinforcement learning algorithms. This innovative method not only tackles storage constraints but also proactively adapts to varying network conditions and traffic patterns. In our approach, we first perform a comprehensive analysis of storage capacities across all nodes, establishing a baseline for dynamic resource allocation. The DLU-RL framework then swiftly assigns storage space based on real-time demand and priority, optimizing storage utilization on the fly. As a result of implementing DLU-RL, substantial enhancements in throughput and concurrent minimization of end-to-end delays are achieved. This research not only contributes to efficient storage allocation techniques but also pioneers the integration of reinforcement learning for wireless communication network performance optimization. The proposed framework signifies a paradigm shift in storage management, offering adaptability, efficiency, and real-time optimization to tackle the evolving challenges of wireless communication.