Performance Improvement of AODV in Wireless Networks using Reinforcement Learning Algorithms

This paper investigates the application of reinforcement learning (RL) techniques to enhance performance Ad hoc On-Demand Distance Vector (AODV) routing protocol in mobile ad networks (MANETs). MANETs are self-configuring consisting nodes that communicate without need for a centralized infrastructure. AODV is widely used due its reactive nature, which reduces overhead and conserves energy. research explores three popular Reinforcement Learning algorithms: SARSA, Q-Learning Deep Q-Network (DQN) optimize protocol's decisions. The RL agents trained learn optimal paths by interacting with network environment, considering factors such as link quality, node mobility, traffic load. experiments conducted using simulators evaluate improvements achieved proposed RL-based enhancements. results demonstrate significant enhancements various metrics, including reduced end-to-end delay, increased packet delivery ratio, improved throughput. Furthermore, approaches exhibit adaptability dynamic conditions, ensuring efficient even highly unpredictable MANET scenarios. study offers valuable insights into harnessing improving efficiency reliability protocols networks.