为了在保证测试精度的前提下提高模拟测试生成算法测试效率,提出了一种基于极限学习机的模拟测试生成算法.首先利用极限学习机对采集的脉冲响应空间样本进行分类,从而减少分类时间消耗,然后联立求解极限学习机表达式和模拟测试生成算法测试序列生成式,得到基于极限学习机的测试序列生成算法表达式,从而简化了测试结构.实验结果表明,基于极限学习机的模拟电路测试生成算法,在保证精度的前提下,可以较好地减少测试时间消耗,提高测试效率.
In order to improve the efficiency and guarantee the precision of test generation algorithm for analog circuit, a extreme learning machine based test generation algorithm is proposed in this paper. Firstly, this algorithm saves time efficiently by using extreme learning machine to classify the pulse response space; Secondly a new process of test signal generator and structure of test in test generation algorithm are presented; and finally, the abovementioned improvement is confirmed in experiments.