Artificial intelligence-based prediction of shear modulus and damping ratio of recycled tire rubber–soil mixtures for sustainable engineering applications

Frugal innovation stands as an imperative cog in the wheel of sustainable development. In the pursuit of simplicity, cost-effectiveness, and environmental compatibility, waste tire rubber and polyurethane-coated rubber (PUcR) emerge as pivotal components in sustainable practices. These materials are advocated for various purposes, including protecting utility tunnels, serving as railway subgrades, and enhancing structural resilience through geotechnical seismic isolation (GSI). Their inherent characteristics, such as modest shear modulus ( G ) and robust damping ratio ( D ), make them well-suited for such endeavors, contributing to sustainability goals by repurposing substantial quantities of non-biodegradable waste. For practicality, leveraging artificial intelligence (AI)-based modern computing techniques for recycled material applications is imperative. In this regard, gene expression programming (GEP) was utilized to develop models for predicting the G and D of rubber–soil mixtures (RSMs) and polyurethane-coated RSMs (PUcRSMs). Employing laboratory testing data from 63 samples across three soil types, the newly proposed models demonstrated exceptional accuracy, with correlation coefficient ( R 2 ) values of 0.91 and 0.97 for G -prediction of RSM and PUcRSM, and 0.9 and 0.86 for D -prediction, respectively. Using AI-based methods, such as GEP to predict mixtures’ dynamic response can cut laboratory costs and optimize mix designs, thereby advancing sustainable material applications.