Decentralized Adaptive Routing Framework using Q-Learning and Blockchain

Efficient and reliable routing in decentralized networks remains a persistent challenge, particularly in environments characterized by dynamic and unpredictable conditions. Traditional routing protocols often lack the adaptability required to respond effectively to rapid changes, leading to suboptimal performance and increased latencies [1] [2]. A decentralized routing framework is proposed, integrating Proof-of-Authority (PoA) blockchain for secure and transparent transaction logging [3] with a multi-neighbor Q-learning algorithm for adaptive decision-making [4] [5]. By leveraging Q-learning, nodes dynamically optimize their routing choices based on latency feedback while balancing exploration and exploitation [6]. The framework is evaluated within a "2-step ring" topology, demonstrating its ability to adaptively select neighbors while maintaining security and decentralization [7]. Results illustrate significant improvements in routing efficiency, adaptability, and decision robustness across multiple operational cycles, offering a scalable solution for modern network challenges [8] [9]