Predictive System-Level Constraint Verification and Optimization

Further automation of analog and mixed-signal integrated circuit design requires the consistent consideration of a growing number of design constraints through all design stages. However, the verification of system-level constraints is only possible towards the end of the design process when all necessary parameters are known. In this paper, we present a method for constraint state prediction in the early stages of an analog IC design project. This allows constraint consideration already during system-level design. By modeling yet unknown design parameters as random variables, the probability of a constraint to be fulfilled can be estimated. Constraint sensitivity analysis is used to identify design parameters with the most influence on a constraint's state. Finally, design parameters are optimized to maximize the probability of fulfilling all constraints.