Topology-Aware Transmission Power Optimization With Graph Convolutional Network for Heterogeneous Wireless Blockchain

In this letter, we investigate the transmission power allocation for heterogeneous nodes in the wireless blockchain network. We analyze the impacts of hash rate, wireless connection bandwidth, and transmission power on forking rate and energy consumption. Based on this analysis, we present a utility model for wireless blockchain nodes. Considering the dynamic topologies of wireless networks and the need to balance blockchain performance with energy efficiency, we propose a Graph Convolutional Network (GCN)-based method to quickly determine power allocation strategies for heterogeneous nodes. Simulation results demonstrate that the proposed GCN method achieves 0.68% Mean Absolute Percentage Error (MAPE) of the classical optimization algorithm and is 85 times faster in terms of runtime. In the dynamic wireless blockchain network, our GCN method decreases the average forking rate by 11.65% compared to the classical optimization algorithm and by 23.68% compared to the random power allocation strategy.