Research on Improvement of Blockchain DPOS Consensus Mechanism Based on HK Clustering

Aiming at the problem of low voting enthusiasm of nodes in the Delegated Proof of Stake (DPOS) consensus mechanism in the blockchain, an incentive mechanism based on a clustering algorithm is proposed. By establishing a node positivity model, the nodes are classified by a clustering algorithm, and rewards are distributed to different nodes according to the classification results, so as to improve the enthusiasm of nodes in voting. The traditional K-means clustering algorithm has problems such as the need to predict the number of clusters K and the inability to deal with outliers, so it is not suitable for the incentive mechanism of node classification. Aiming at the problem of K-means clustering algorithm, an Hegselmann-Krause (HK) opinion dynamics clustering algorithm is introduced, which can meet the needs of the reward distribution mechanism in this paper. Through simulation and comparison with the existing schemes, the effectiveness of the improved algorithm is verified.