基于蚁群和遗传算法的测试向量生成方法

A Method to Generate Verification Vector Based on Ant Algorithm and Genetic Algorithm

随着设计复杂度的迅速增长,集成电路的测试已成为阻碍其发展的重要因素,如何尽可能自动生成可以满足测试覆盖率的测试向量是这一问题的关键所在。本文在对测试向量自动生成问题分析的基础上,建立了数学模型,并提出了一种适合求解该问题的蚁群遗传融合优化算法。该方法首先由蚁群算法得到测试向量集,然后利用遗传算法对向量集进行优化。实验数据表明,通过该算法,只需较少的迭代次数就可以自动生成满足一定覆盖率的测试向量组,由此可以证明该方法在产生高覆盖率测试向量上具有一定的有效性。

As the increase of the difficulty during an ASIC design, verification has become an important problem in the development of Integrate Circuit. The key point to settle this problem is how to generate verification factors satisfying the coverage as automatically as possible. Based on the analysis about this problem, a mathematical model is built and a combined algorithm of ant algorithm and genetic algorithm is designed. This method adopts ant algorithm to generate a class of verification factors and then uses genetic algorithm to optimize the class. Experimental results show that verification factors based on high coverage will be generated automatically only after several iterative times and this method is effective and efficient.