Unified Post-Quantum Cryptography For Cloud-Assisted Iot: Lattice Homomorphic Proxy Re-Encryption With Ghoa-Optimized Parameters and Explainable Security

Abstract

Medical and industrial IoT deployments require cryptographic solutions that provide confidentiality for data being computed in the cloud, enable secure delegation to authorized users without disclosing private keys, and remain quantum resilient under the energy and memory constraints of ARM Cortex-class devices. Current methods only partially address these requirements. This proposal arises from the observation that homomorphic encryption schemes can be constructed with a high level of abstraction. The result is a hybrid that selectively supports specific operations through a cryptographic transformation. The Operation homomorphic is categorized into several types. ALPHREN synthesizes four synergistic elements. The multi-objective Ring-LWE parameter optimization problem (which considers security, encryption latency, and memory consumption) is tackled with a Genetic Honey Optimization Algorithm (GHOA). The GHOA converges to around 47 iterations and an optimal parameter set (n = 1024, q = 12289, σ = 3.19). The Batch Fan-Vercauteren homomorphic encryption engine can carry out operations on a depth d ≤ 7 arithmetic circuits whilst allowing for a noise growth that is correctable. One more, a method of one-way proxy re-encryption reliably changes ciphertexts between users without disclosing plaintext. Ultimately, a SHAP (SHapley Additive exPlanations) module interprets the contribution of each parameter to the security-performance trade-off to enable HIPAA and GDPR compliant transparency. We ran the framework on Raspberry Pi 4B and used the datasets MIMIC-III (48,321 clinical records) and UCI IoT Intrusion (625,783 records) for 30 independent trials. ALPHREN was able to generate the keys in 0.43 ms, achieve an encryption throughput of 2.14 MB/s and an energy consumption of 4.7 μJ per operation with a security classification accuracy of 99.1%. Encryption latency was 32.8× lower than the closest PQC competitor. Both the Friedman test (χ²(6) = 44.19, p < 0.001) and Bonferroni-corrected Wilcoxon tests showed these results to be statistically significant and of large effect size (Cohen’s d = 1.39–2.11). ALPHREN is the first framework that integrates post-quantum security, homomorphic evaluation, proxy re-encryption and explainable cryptographic parameter optimization on resource-constrained IoT hardware.