论文部分内容阅读
面对集成电路(IC)在寿命试验中难以得到大量失效数据的小子样情形,提出了基于最小二乘支持向量机(LSSVM)的小子样正态寿命型IC可靠性评估方法。该方法的主要思想是基于寿命试验数据建立最小二乘支持向量机回归模型,根据该模型计算出正态分布的参数,从而进行可靠性评估。用蒙特卡罗方法研究了截尾失效情况下该方法在正态寿命型IC平均寿命评估应用中的可行性,同时与常用的最小二乘回归(LSR)法和极大似然估计(MLE)法相比,结果表明,基于LSSVM的方法能更精确地反映小子样下IC的可靠性,能为评估小子样正态寿命型IC的可靠性提供一种新的有效途径。
In the face of a small sample of ICs that is hard to get a large number of failure data in the life test, a small sample normal life IC reliability evaluation method based on least squares support vector machine (LSSVM) is proposed. The main idea of this method is to establish the regression model of least square support vector machine based on the life test data, and to calculate the parameter of normal distribution according to the model so as to evaluate the reliability. Monte Carlo method is used to study the feasibility of this method in the application of life expectancy to the average life expectancy of IC under the condition of tail loss. Meanwhile, the results are compared with the commonly used least squares regression (LSR) and maximum likelihood estimation (MLE) The results show that the LSSVM-based method can reflect the reliability of the IC under the small sample more accurately and provide a new effective way to evaluate the reliability of the small sample-like normal life IC.