Performance Evaluation of Two Allocation Schemes for Combinatorial Group Testing Fault Isolation

Two fault isolation approaches based on Combinatorial Group Testing (CGT) are presented. Although they both share the basic principle of grouping suspect resources into subgroups for testing, they differ in the allocation strategy used to achieve that grouping. Equal share CGT approach is analyzed first, results shown successful fault isolation in 85% of the runs. An average of 3.65 groups and 362 test vectors were needed to isolate a single fault with an average of 2.95 discrepant outputs in each run. After that, Interleaved CGT approach is proposed along with analytical demonstration of its strengths and weaknesses. Finally, future research framework is suggested to experimentally compare the two