Modeling of MOS transistors based on genetic algorithm and simulated annealing

A novel method to extract an efficient model for metal-oxide-semiconductor (MOS) transistors in order to satisfy a specific accuracy is presented. The approach presented here utilizes a genetic algorithm (GA) to choose the necessary physical and heuristic elements in order to define a compact yet accurate model for MOS I-V characteristic. Then the values of the free parameters related to each element are determined using simulated annealing (SA). For a desired accuracy considered here, the accuracy of the results predicted by our model were within 3.1%, for PMOS, and 1.3%, for NMOS, of the results of BSIM3 model while having much less complexity compared to the BSIM3 model. When this model with a variable accuracy is implemented in a circuit simulator, it provides the freedom of making a selection between the time and the accuracy of the simulation.