OPTIMA - A NONLINEAR MODEL PARAMETER EXTRACTION PROGRAM WITH STATISTICAL CONFIDENCE REGION ALGORITHMS

This paper describes a device model parameter optimization program, OP TIMA, for extracting parameter values of empirical and/or analytical m odels most commonly used for VLSI circuit simulation. Such device mode ls often have parameters which are correlated, and some of them could be redundant. The program OPTIMA can automatically detect redundant mo del parameter combinations. This not only allows one to extract more m eaningful parameters, but also helps in the development of improved ph ysical models with a minimum number of empirical parameters. The param eter redundancy is detected using a statistical confidence region algo rithm which can be implemented as a post-processor to any gradient-bas ed least squares optimization method. The advantage of the statistical confidence region algorithm in OPTIMA, as applied to MOSFET model par ameter extraction, is discussed using examples from drain and substrat e current modeling.