@Article{cmes.2026.084699,
AUTHOR = {Muammar Shahrear Famous, Samia Sayed, Rashed Mazumder, Risala T. Khan, M. Shamim Kaiser, Mohammad Shahadat Hossain, Karl Andersson},
TITLE = {Explainable AI (XAI)-Based Security Verification in Blockchain-Enabled Drug Supply Chains},
JOURNAL = {Computer Modeling in Engineering \& Sciences},
VOLUME = {},
YEAR = {},
NUMBER = {},
PAGES = {{pages}},
URL = {http://www.techscience.com/CMES/online/detail/28282},
ISSN = {1526-1506},
ABSTRACT = {Counterfeit pharmaceuticals, fragmented data management, and the absence of transparent verification mechanisms continue to threaten the integrity of modern healthcare drug supply chains. Although blockchain-based traceability frameworks improve decentralization and auditability, existing approaches frequently lack formally verified authentication guarantees and interpretable security decision mechanisms. To address these limitations, this paper proposes a unified blockchain-enabled healthcare supply-chain framework integrating nonce-based cryptographic authentication, formal security verification, Solidity smart contracts, and SHAP-based explainable artificial intelligence (XAI). The proposed protocol employs elliptic-curve digital signatures, nonce-based challenge–response authentication, and blockchain validation mechanisms to ensure secure communication among manufacturers, distributors, pharmacies, and regulatory authorities. Formal verification using ProVerif under the Dolev–Yao adversarial model confirms secrecy, authentication, integrity, and replay-attack resistance properties, with no executable attack traces detected. A permissioned Ethereum test network and Python-based cryptographic simulation environment were implemented using Solidity, Web3.py, and Py-EVM to evaluate deployment feasibility. Additional deployment validation was performed through Ethereum Sepolia test-network execution and Mythril smart-contract security auditing. Experimental results demonstrate stable and scalable operational performance, achieving average transaction latency below 3 s, throughput performance comparable to recent healthcare blockchain systems reporting approximately 95–155 transactions per second, and average cryptographic signing time of 0.71 ms per transaction. Gas-consumption analysis reported a mean transaction cost of 182,345 gas units with low execution variance, while storage overhead exhibited near-linear scalability as transaction volume increased. Additional adversarial evaluation demonstrated resilient transaction-validation behavior under varying node participation and simulated communication disruptions. To improve interpretability and regulatory transparency, the framework integrates SHAP-based explainable verification capable of identifying the relative influence of signature validity, nonce freshness, blockchain confirmation, and integrity verification on transaction acceptance decisions. Ablation analysis additionally showed that the complete integrated framework achieved 96.8% security assurance, 94.5% interpretability, and 97.2% traceability performance, substantially outperforming partial configurations lacking blockchain, formal verification, or explainability modules. The principal contribution of this work lies in combining formally verified authentication, blockchain-enabled traceability, explainable security analytics, and deployment-oriented validation within a unified pharmaceutical supply-chain architecture suitable for secure, scalable, and regulation-aware healthcare environments.},
DOI = {10.32604/cmes.2026.084699}
}