Using Non-uniform Crossover in Genetic Algorithm Methods to Speed up the Generation of Test Patterns for Sequential Circuits

  • Authors:
  • Michael Dimopoulos;Panagiotis Linardis

  • Affiliations:
  • -;-

  • Venue:
  • SETN '02 Proceedings of the Second Hellenic Conference on AI: Methods and Applications of Artificial Intelligence
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

Due to the high complexity of the problem of generating test patterns for digital circuits Genetic Algorithms (GA) have been investigated as an alternative to deterministic algorithms for test generation. In this paper a Genetic Algorithm "GATPG" is presented for generating sequences of test vectors for sequential circuits. The aim is to produce compact test sequences that attain high fault coverage. Because of the constraints imposed on a GA by the peculiar characteristics of sequential circuits it is proposed here a non-uniform selection probability for crossover combined with individuals (test sequences) of variable length and a two-phase fitness function. For the evaluation of candidate test sequences is used a 3-valued fault simulator, allowing the test patterns to be applied on faulty circuits that start from an arbitrary (unknown) state. Experimental results with respect to the ISCAS'89 benchmarks are presented to show the viability of the proposed approach.