Dynamic optimization of encryption parameters in private blockchain networks using computational intelligence

This study aims to develop and experimentally validate an adaptive encryption system for private blockchain environments using computational intelligence for dynamic parameter optimization. The proposed framework integrates symmetric and asymmetric encryption, PBFT-based consensus, and a regression-based load prediction model that dynamically adjusts key parameters and block characteristics. A mathematical model linking message length, algorithmic complexity, processing time, and verification latency enables quantitative performance evaluation. Experimental results show that the adaptive approach reduces average cryptographic processing time by 59.8%, decreases overall transaction latency by 41.3%, and improves network throughput by 61.6% compared to static configurations while maintaining security constraints. The findings confirm the effectiveness of integrating computational intelligence with blockchain architectures to achieve a balanced trade-off between security and performance. Additional experiments across 4–13 PBFT nodes and a heterogeneous seven-node configuration showed that the adaptive performance advantage persisted, although increasing consensus overhead reduced the relative latency benefit in larger networks.