A method for dynamic account allocation in a sharded blockchain considering transaction type
By optimizing account allocation in a sharded blockchain system using graph attention networks and deep reinforcement learning algorithms, the problem of uneven resource consumption caused by differences in transaction types is solved, thereby improving the system's scalability and performance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-25
- Publication Date
- 2026-07-24
AI Technical Summary
Existing sharded blockchain systems fail to effectively consider differences in transaction types when allocating accounts, resulting in uneven resource consumption, difficulty in coping with complex and ever-changing transaction patterns, and impact on system scalability and performance.
By employing graph attention networks and deep reinforcement learning algorithms, combined with transaction type weights, an account allocation strategy is optimized. The account migration process is modeled through a GA-PPO agent to achieve dynamic account allocation.
It improves the scalability and performance of sharded blockchain systems, optimizes resource utilization, and reduces cross-shard transaction rates and system load imbalance.
Smart Images

Figure CN120931284B_ABST
Abstract
Citation Information
Patent Citations
Inductive block chain account distribution method and device based on graph neural network
CN117391858A
Fragmented block chain account distribution method and system
CN119322672A