A Proposal of Genetic Operations for BSIM Parameter Extraction Using Real-Coded Genetic Algorithm

This paper discusses genetic operations and their effects on evolution of GA in BSIM parameter extraction problems. Generally, Real-Coded Genetic Algorithm (RCGA) using Simplex Crossover (SPX) is often employed to extract BSIM parameter sets. BSIM parameters, however, have recommended operating ranges. There are regarded as constraints, thus all extracted parameters have to be satisfied them. In many cases, when the number of parameters becomes large, the conventional methods generate a lot of infeasible solutions because SPX makes offspring on the simplex plane expanded by ε parameter. This makes search efficiency of GA reduce drastically. Because of these factors, we propose genetic operations considering the constraints to prevent reduction of search efficiency of GA. In this paper, some experiments using actual static characteristic curves of MOS-FET were conducted to validate the proposed method. This paper also discussed the effectiveness of the proposed method.