THE SECOND SAUDI INTERNATIONAL CONFERENCE ON NUCLEAR POWER ENGINEERING (SCOPE-2)

The safe decontamination of abandoned radioisotope and radiopharmaceutical production facilities in Indonesia for reuse is a pressing need due to the increasing local demand and the country's current dependence on imports. This study presents a structured risk evaluation model for the decontamination process in such facilities, integrating qualitative risk assessment with the Analytic Hierarchy Process (AHP) as a deterministic machine learning approach. The model prioritizes risk control strategies based on five critical criteria: effectiveness, cost, implementation time, regulatory compliance, and environmental impact. AHP was employed to calculate the relative weights of each criterion and evaluate four alternative risk control measures through pairwise comparisons and eigenvector calculations. The results identified radiation shielding (Alternative A1) as the most effective control measure, scoring 0.3925, significantly outperforming the other alternatives. This prioritization framework reflects a machine learning-aligned decision-making process, enabling data-structured reasoning for safety and regulatory adherence. While the current method is deterministic, its architecture supports the future integration of dynamic, data-driven machine learning models that are capable of real-time risk prediction and adaptive control strategy. Overcoming current limitations, such as static input assumptions and limited operational feedback, could pave the way for intelligent, automated decision-support systems that enhance radiation protection and operational efficiency in high-risk environments.