Bayesian Optimization for Stack-up Design

Abstract

Black-box function optimization is a challenging problem worldwide. Bayesian Optimization is a powerful method used to handle the optimization of functions, which are usually too costly for evaluation. Non-convex black-box function optimization arises in many applied problems. One example of such kind of problems is the PCB stack-up design. With its various covariance functions and different values of hyper-parameters Bayesian Optimization is applied for five-dimensional stack-up design optimization. The results obtained using the Bayesian Optimization are compared with results of other methods.