Design centering of analog circuit's component values using parallel genetic algorithms

Genetic algorithms technique for analog circuit design automation tool has previously been described. This paper shows preliminary results for the application of parallel genetic algorithms to solve the problem of design centering by means of providing an estimate of the feasible region of a circuit design problem. This is achieved by performing a two-stage GA search. In the first stage a general search is conducted over the whole solution space to identify regions of feasibility. While in the second stage the search focuses on these estimated feasible regions.