Distributed Deep Reinforcement Learning-Based Content Caching in Edge Computing-Enabled Blockchain Networks

Abstract

As an emerging technology, blockchain possesses the decentralized trust establishment, preservation and delivery capabilities, but its application is limited by users' computing and storage capabilities. With the aid of Mobile edge computing (MEC) technology, frequently used block data of blockchain network can be cached in the edge servers to improve communication efficiency and reduce data transmission delay. In this paper, we investigate the problem of content caching in MEC-enabled blockchain network. In order to encourage edge caching and reduce transmission delay with high efficiency, we adopt distributed deep reinforcement learning (DRL) in solving the high-dimensional and large-scale decision-making problem. The simulation results show that the proposed scheme based on distributed proximal policy optimization (DPPO) outperforms the schemes based on other existing DRL algorithms.