Machine Learning Based Modeling of Electrical Characteristics in Triangular Gate FinFETs for Low Power Electronics

Abstract

Modeling and optimization of devices play a critical role in the management of product quality and the advancement of technology within the industrial sector. With the advent of novel devices and the progression of technology, these devices exhibit a multitude of interrelated factors and demonstrate a nonlinear correlation. Triangular Gate (TG) FinFETs technology has emerged as a possible alternative for addressing the limitations of traditional planar transistors in present integrated circuits (ICs). This paper presents an effective data-driven Multiobjective Optimization (MOO) with evolutionary computation (EC) techniques. By using these techniques, TG FinFETs enables the automated identification of optimal design that balances the transistor speed, power, and variability. To assist in the design of TG FinFETs, this study integrated two popular MOO techniques such as PAL and NSGA-III. These algorithms effectively handle the complicated trade-offs between diverse objectives and allow for efficient and effective TG FinFETs design optimization.