Application of artificial intelligence techniques in modeling attenuation behavior of ionization radiation: a review

Introduction Shielding of ionizing radiations, which are gamma rays, neutrons, and X-rays, can be achieved by attenuating its intensity using different materials. Protection is therefore crucial in ensuring the safety of lives and essential equipment in areas such as nuclear power plants, radiotherapy facilities, space exploration, and others. Artificial Intelligent technologies have become desirable in modeling shielding materials’ attenuation behavior due to their unique advantages. Objective The overview aims to present the recent application of AI technologies in modeling the radiation attenuation behavior of materials. Methods A total of 41 relevant articles were obtained using Scopus and web of science databases. The search was restricted to articles and conference papers published within the last two decades. Results From the overview, it was realized that AI techniques can predict the attenuation properties of shielding materials and optimize the shield design. The methods can be grouped into predictive models which are: fuzzy logic, Support Vector Regression, Neural Networks, and optimization models which include Genetic algorithms, Ant Colony, and Particle Swarm Optimization. Neural networks are the most robust and widely used technique. The predictive models are used in predicting parameters such as attenuation coefficient, buildup factor, shield thickness, and radiation dose rates, whiles the optimization techniques are employed in single and multi-objective attenuator designs. Conclusion In the overview, the accuracies and complexities of the various AI techniques have been discussed giving insight into their prospects. The AI techniques are easy to model compared to conventional methods and can save computational time when coupled with conventional statistical and deterministic models or employed as a standalone technique.