RaGAzi: A random and gradient-based approach to analog sizing for mixed discrete and continuous parameters

Sizing of analog circuits requires the consideration of both continuous and discrete design parameters, e.g. due to predefined manufacturing grids. In this paper, a new method is presented to solve this task. It is based on an iterative optimization process with random-based searches on search regions that are determined by performance gradients. Experimental simulation results of operational amplifiers show the efficiency and efficacy of the presented method.