The Merge changes Ethereum from Proof-of-Work (PoW) to the more secure and less energy-intensive Proof-of-Stake (PoS) mechanism. However, the existence of malicious validators still threatens the security of Ethereum,...
详细信息
ISBN:
(纸本)9798350377859;9798350377842
The Merge changes Ethereum from Proof-of-Work (PoW) to the more secure and less energy-intensive Proof-of-Stake (PoS) mechanism. However, the existence of malicious validators still threatens the security of Ethereum, primarily through a discouragement attack. How can we redesign the incentive mechanism in PoS Ethereum for a more secure blockchain? For this quest, we, for the first time, apply the cutting-edge reinforcement mechanism design method-an interdisciplinary approach at the intersection of reinforcement learning (RL) and mechanism design-to staking mechanism designs. We abstract a generalized staking mechanism as a game environment and implement an RL method for the blockchain as a mechanism designer to explore the optimal incentive design. Our reinforcement mechanism design outperforms the status quo in cultivating honest validators. Furthermore, we identify Advantage Actor-Critic (A2C) as the most efficient RL algorithm among the three alternatives, which intuitively performs better when the initial proportion of honest validator is larger. Our interdisciplinary approach of generalized abstraction could be adapted to analyze the incentive design in any PoS blockchain and beyond.
暂无评论