Blockchain based Single Path and Multi-Path Routing in Payment Channel Network using Artificial Intelligence Techniques

Blockchain-based Payment Channel Network (PCN) routing for payment transactions includes the transfer of funds between multiple parties through a network of interconnected payment channels. These channels permit off-chain transactions, balance payment channels, enhance speed, and reduce the transaction costs. However, finding the best route for transactions between parties is challenging due to the dynamic nature of the network. The routing concept for PCN is divided as single-path routing and multi-path routing to enhance transactions based on the requirements of the users/organizations. Channel-based strategy, off-chain transactions, channel rebalancing techniques, and payment execution methods are considered essential in PCN transactions for an efficient routing. This study analyses the following methodologies: Deep Neural Network (DNN), Deep Reinforcement Learning (DRL), Graph NN (GNN), as well as genetic optimization algorithm utilized for efficient routing. Currently, distance-based methods like Markov chain-based routing are also being employed in PCN for finding an optimal path for payments. This survey examines the strengths and challenges of these Deep Reinforcement learning-based routing approaches.