Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
545713 | Microelectronics Journal | 2014 | 9 Pages |
Once all available measurements are determined, the highest testability index of a complex system is determined. To achieve such highest index with the lowest test cost, AND/OR graph search algorithms were developed for years to determine an optimal or near-optimal test sequence. However, in most cases, achieving the highest testability index induces extremely high test cost. The purpose of this paper is to optimize test set and test sequence so as to cut down the test cost while keeping the required, not necessarily the highest, FIR (Fault Isolation Rate) satisfied. Traditionally, this is an NP-Complete problem, which makes the computation of optimal test set impractical for even the moderate-sized model. In this paper, a greedy method is proposed to get the optimized test set. Then, we combine the greedy method with discrete binary particle swarm optimization (DPSO) to construct a test sequential tree. With the specified FIR requirement satisfied, the lowest test cost is achieved. The proposed algorithms are illustrated and tested in a range of real-world systems. The effectiveness and accurateness of the proposed algorithm is verified by computational results.