Testability measures combining probablistic and sampling techniques

Probabilistic methodologies and sampling techniques are alternate approaches to the problem of determining quantitative testability measures. Unfortunately, both methods are not sufficiently accurate in treating hard-to-detect faults. An approach in which probabilistic methodologies are used to improve the effectiveness of testability measures obtained by means of the simulation of a few random vectors is presented. Results obtained on ISCAS benchmarks show relevant improvements on the accuracy in the evaluation of fault detection probabilities.<
>