FPAD: a fuzzy nonlinear programming approach to analog circuit design

FPAD, a CAD tool for the design of analog circuits, is presented. For a set of input specifications, FPAD deduces the device sizes that optimize the performance objectives while satisfying the constraint specifications. Fuzzy set theory is used to measure the degree of fulfilment of the objectives and constraints, and hence provides a measure of the design quality. Trade-offs are handled by manipulating the shape of the membership functions that reflect the fulfilment or violation of the performance specifications. FPAD allows the optimization of several objectives simultaneously since multiple objective optimization is used. The mathematical formulation of the design problem is fuzzified to support real-world terms like high, good, small, acceptable, etc. The optimization is based on analytic models to avoid a simulator in the inner optimization loop. Using this approach, a two-stage unbuffered CMOS op amp and a simple emitter follower were designed, with promising results.<
>