Learning-Based Off-Chain Transaction Scheduling in Prioritized Payment Channel Networks

【Author】 Luo, Xiaofei; Li, Peng

【Source】IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS

【影响因子】13.081

【Abstract】Payment channel network (PCN) is one of the promising solutions for scalable blockchains since it shows great potential in improving blockchain network throughput. However, the growing number of transactions and the payment-channel sharing of concurrent transactions can lead to channel congestion. Although many studies have proposed different solutions to solve this problem, they ignore a fact that applications may have different transaction rate requirements at different times. In this paper, we propose a priority-aware PCN to meet the requirements of those transactions. Senders in priority-aware PCNs can specify the priority of their transactions by paying a corresponding forwarding fee on each hop along the transaction path. However, capacity competition occurs on the shared hops. Moreover, we propose a multi-agent DQN-based priority assignment algorithm to address the competition issue and design a PCN simulator for performance evaluation. Simulation results show that our solution can guarantee a high throughput of transactions and assign priorities appropriately to balance the transaction rate and forwarding fee cost. The experimental results demonstrate that the priority scheduling scheme can achieve higher transaction throughput and success ratio than other scheduling methods in a congested PCN environment.

【Keywords】Payment channel network; off-chain transaction scheduling; priority; reinforcement learning

【发表时间】2022 DEC

【收录时间】2023-01-14

【文献类型】实验仿真

【主题类别】

区块链技术-协同技术-支付通道网络