Genetic algorithm based extraction of IC device model parameters

As genetic algorithms (GAs) have high robustness in finding the global optimum of functions with multi-extremum no matter what the starting points are, yet the traditional search methods are easy to be trapped into the local optima if the initial points are not good enough. GAs are applied to extracting model parameters of semiconductor devices in this paper. After addressing the main concept of GAs and their superiority to traditional optimization algorithms, some advanced strategies and knowledge-based techniques are proposed to improve the global convergence and efficiency of solving the problem of the model parameter extraction (MPE). As it proves that the DC model parameters of bipolar junction transistors (BJT) can be successfully extracted by the GA-base algorithms, the methodologies proposed in this paper may be extended to MPE problems of other IC devices