The Journal of China Universities of Posts and Telecommunications ›› 2021, Vol. 28 ›› Issue (2): 68-78.doi: 10.19682/j.cnki.1005-8885.2021.1006

Previous Articles     Next Articles

Performance optimization for smart grid blockchain integrated with fog computing using DDQN

Xue Chenzi, Wei Yifei, Zhang Yong   

  1. School of Electronic Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2020-08-25 Revised:2021-03-06 Online:2021-04-30 Published:2021-04-30
  • Contact: Xue Chenzi

Abstract: In order to solve the energy crisis and pollution problems, smart grid is widely used. However, there are many
challenges such as the management of distributed energy during the construction. Blockchain, as an emerging
technology, can provide a secure and transparent solution to the decentralized network. Meanwhile, fog computing
network is considered to avoid the high deployment cost. The edge servers have abundant computing and storage
resources to perform as nodes in grid blockchain. In this paper, an innovative structure of smart grid blockchain
integrated with fog computing are proposed. And a new consensus mechanism called scalable proof of cryptographic
selection (SPoCS) is designed to adapt the hybrid networks. The mechanism not only includes a special index,
contribution degree, to measure the loyalty of fog nodes and the probability of being a function node, but also has
flexible block interval adjustment method. Meanwhile, the number of function nodes (validating nodes and ordering
nodes) can also be adjusted. And a deep reinforcement learning (DRL) method is used to select the appropriate
quantity to improve the performance under the strict constraints of security and decentralization. The simulation
shows the scheme performs well in the throughput, cost and latency.

Key words: blockchain, smart grid, fog computing, consensus, deep reinforcement learning (DRL)

CLC Number: