Multi-Objective Optimization Algorithms for Automated Circuit Sizing of Analog/ Mixed-Signal Circuits

Manual circuit design involves a time-consuming sizing task that diverts the highly skilled analog designers from creative work to a trial-and-error process. To reduce R&D costs and speed-up the time to market, this process can be automatized using artificial intelligence. In this paper we evaluate the most promising five Evolutionary Algorithms in the context of circuit sizing in a multi-objective optimization scenario. We calibrate their hyperparameters by performing a grid search on synthetic benchmarks and we compare the performance of the algorithms on a real voltage regulator.