Using Tabu Search for optimization of memory-constrained hybrid BIST

Abstract

This paper deals with optimization of hybrid BIST testing approach with memory constraints. The traditional external tester is often unfeasible for embedded systems and therefore different self-test solutions are sought after. Classical built-in self-test (BIST) approaches are largely based on pseudorandom testing and using linear feedback shift registers (LFSR). One of the possible extensions of classical BIST is hybrid BIST, where pseudorandom tests are complemented with precomputed deterministic test patterns to increase test coverage and reduce test time. Hybrid BIST optimization method based on Tabu Search is proposed in this paper. As a generalization of local optimization, Tabu search method is used for finding the optimal solution. The objective is to minimize test length under given memory constraints, without losing test quality.