一种GP参数优化的有效方法

AN EFFECTIVE METHOD FOR PARAMETER OPTIMIZATION OF GP MODEL

本文提出了一种双极集成电路GP模型参数优化的有效方法.该算法对GP模型迭代方程的参数和节点电压初值的要求大为降低,同时代替非线性方程的求解,以确定本征结电压的多次迭代过程.另外,对优化过程中所采用的简单变换处理箱型约束的方法进行了分析,确定所存在问题并在优化过程中采用了Watchdog方法使这类约束问题得到方便地解决.实验表明,这些处理对GP参数的优化是极为有效的.

An effective method for optimal parameter extraction of GP models in bipolar integrated circuits is suggested, Emphasis is placed on processing the iterative equations of GP models, The purpose is to reduce the requirements of the algorithm for initial model parameters and node voltages, And the iterative procedures of solving nonlinear equations for determining intrinsic junction voltages are replaced by an improved method, The method that box constraints are transformed simply is analyzed, A watchdog algorithm is tised to find the solution of the constraint problem, Experiments show that the method is very effective for optimal parameters of GP models,