A Machine Learning based Metaheuristic Technique for Decoupling Capacitor Optimization

Abstract

Decoupling capacitors are commonly used in the design and optimization of Power Delivery Networks (PDNs) in high-speed very large scale integration systems (VLSI) to minimize the variations in the power supply and to maintain a low PDN ratio. In this paper, an efficient and fast Machine Learning (ML) based surrogate-assisted metaheuristic approach is proposed for the decoupling capacitor optimization problem to reduce the cumulative impedance of the PDN below the target impedance. The performance comparison of the proposed approach with state-of-the-art approaches is also presented.