WEMA25: A Machine Learning Generative AI Approach for Antenna Design and Topology Optimization

Presentation Abstract
Conventional antenna design workflows require deep domain knowledge for topology selection. Additionally, antenna design requirements are complex and changing across different applications areas. In this MicroApp, we present Cadence Optimality Intelligent System Explorer's AlphaGo algorithm approach to antenna design problems as a kind of game. Starting from a blank grid, AlphaGo algorithm generates antenna layouts that can meet designers’ specifications. During the generation process, a machine learning model with deep neural network architecture is built and improved, thus facilitating an online optimization flow with an accurate surrogate model that can accelerate design efficiency and reduce design cycles.