Reinforcement Learning for the Optimization of Power Plane Designs in Power Delivery Networks

Abstract

This paper proposes a deep deterministic policy gradient (DDPG) based method to optimize the power plane in power delivery networks (PDNs). The proposed method considers the degrees of freedom of a plane design in a layer, determining the parameters for creating a power plane. The results show that the proposed method can provide an optimized power plane design even in a plane layer with a restricted region.