A Neural Network Modeling Approach to Circuit Optimization and Statistical Design

By using the accurate models such as physics based FETS models coupled with the demand for Yield optimization the result is challenging. A new approach to microwave circuits optimization and statistical design featuring neural network models at either device or circuit level is presented in this paper. A physics based-oriented FET model yet without the need to solve device physics equation repeatedly during optimization is presented by neural network at device level. At circuit level the neural network speeds up optimization by replacing repeated simulation. As comparison to direct optimization of original device and circuit models this method is faster. Compare to existing polynomial or table look-up models used in analysis and optimization the proposed approach has the capability to handle high dimensional and nonlinear problems.