ATPG for Sequential Circuits through Genetic Simulation and Extracted Knowledge Exploitation

A GA based sequential circuit test generator is presented. In addition to the typical GA operations knowledge extracted during the evolutionary process is used to further guide test generation. Population sequences having low fitness values but detect different faults, which otherwise would be discarded, are combined and specially compressed into a single sequence that covers, at least, the faults of each individual sequence. This technique combined with specialized GA operators leads to higher fault coverages at shorter computation times. Experimental results support the usefulness of the proposed method. Key-Words: Sequential Digital Circuits, Test Generation, Genetic Algorithms.