The Huber concept in device modeling, circuit diagnosis and design centering

We present exciting applications of the Huber concept in circuit modeling and optimization. By combining the desirable properties of the l/sub 1/ and l/sub 2/ norms, the Huber function is robust against gross errors and smooth w.r.t. small variations in the data. We extend the Huber concept by introducing a one-sided Huber function tailored to design optimization with upper and lower specifications. We demonstrate the advantages of Huber optimization in the presence of faults, large and small measurement errors, bad starting points and statistical uncertainties. Circuit applications include parameter identification, design optimization, statistical modeling, analog fault location and yield optimization.<<ETX>>