ANN–GA Based Modeling of Mechanical Behavior in Nano Hybrid Biocomposites

Abstract

The increasing demand for lightweight, high-performance, and sustainable materials has accelerated the development of nano-hybrid biocomposites for advanced engineering applications. However, their mechanical behavior is highly nonlinear and strongly influenced by multiple interacting parameters, including fiber content, nanoparticle loading, dispersion quality, and processing conditions. This study presents an integrated Artificial Neural Network (ANN) and Genetic Algorithm (GA) approach to model and optimize the mechanical performance of nano-hybrid biocomposites. Experimental datasets were generated through systematic variation of bio-fiber reinforcement and nanofiller content in a biodegradable polymer matrix. Tensile, flexural, and impact properties were measured and used to train a multilayer feedforward ANN. The trained network demonstrated high predictive accuracy, with correlation coefficients exceeding 0.98. Subsequently, GA was employed to identify optimal material compositions maximizing strength and stiffness while maintaining impact resistance. The hybrid ANN–GA framework effectively captured complex nonlinear relationships and provided reliable optimization of biocomposite formulations. The proposed methodology offers a powerful tool for intelligent material design, reducing experimental cost and supporting the development of sustainable, high-performance composite systems.