This article introduces an advanced difficulty adjustment algorithm for the Ethereum blockchain built on the Proof-of-Work (PoW) mechanism, utilizing deep learning to maintain low volatility in block difficulty and generation time. Simulations with actual data demonstrate that the algorithm based on Long Short-Term Memory (LSTM) networks outperforms other baseline models in maintaining this low volatility. The study indicates that LSTM is more effective in controlling data volatility and can capture the trends in the original test dataset. Despite the limitations in runtime associated with deep learning methods, the research also presents potential approaches to reduce training time through incremental learning and explores the prospects of implementing this method on the Bitcoin chain.
Sixu Di(2024). Enhancing Blockchain Stability with LSTM-Based PoW Difficulty Adjustment: A Deep Learning Approach. Journal of New Economics and Finance, Volume 4, Issue 1, Page 1-12, Publish Date: 2024-01-19
The views expressed in this article are solely those of the author and do not represent the views of the journal or its editors. If you believe there are any inappropriate aspects in the content or viewpoints of the article, please click the "Feedback" button below the article and submit your comments as instructed.
English
Chinese