BIO-INSPIRED PREDICTIVE INSIGHT AND OPTIMIZATION STRATEGY FOR ELECTROMAGNETIC SHIELDING IN HYBRID POLYMER COMPOSITES

Abstract

The Machine learning-based EMI shielding effect prediction on polymer composite materials is of supreme importance in modern electronic applications as this shielding serves as a robust protective barrier, safeguarding electronic components from electromagnetic interference. But, capturing the multifactorial nature of the composites and handling multi-objective optimization is still difficult to achieve with the existing machine-learning approaches. Hence, this research proposes a novel “Bio-inspired Predictive Insight and Optimization Strategy” for improving the efficacy of prediction and optimization of the EMI shielding effect of polymer composites. Polymer composites have complex structures with multiple components, leading to nonlinear relationships with EMI shielding effectiveness. This results in a high-dimensional search space, making the search for effective solutions computationally demanding and time-consuming. Here, the “Bayesian-Enriched Genetic Programming” approach captures the multiple parameters such as filler loading, matrix type, and processing conditions effectively to reduce this high-dimensional space complexity. The “Multi-objective Dominant Crowding Seagull (MDCS) optimization” model effectively addresses the challenges of traditional optimization algorithms by optimizing multiple conflicting objectives and EMI shielding. The comparison results show that the proposed method outperforms other methods with a higher prediction accuracy of 98.7%, faster training time of 19 seconds and quicker prediction time of 6 seconds.