Parametric yield optimization of electronic circuits via improved centers of gravity algorithm

An improved centers of gravity (CG) algorithm is presented for circuit yield optimization or for the well-known design centering (DC) problem. The CG algorithm is based on the estimation of circuit yield and CG of pass and fail sampled points. A new type of estimator and the Latin hypercube sampling (LHS) scheme are proposed in order to improve the estimation accuracy with the same computational time. Simulation results have shown a good agreement with the theoretical basis.