Search strategy switching: A cost model and an analysis of backtracking

Test generation algorithms contain search strategies which are used to control decision making when the algorithm encounters a choice of signal value, or what action to perform next. Our study of traditional search strategies used in automatic test pattern generation has led us to the observation that no single strategy is superior for all faults in a circuit and all circuits. Further experimentation led to the conclusion that a combination of search strategies provides better fault coverage and/or faster ATPG for a given backtrack limit. Instead of using just one strategy up to the backtrack limit, a primary strategy is used for the first half of the backtrack limit, then a secondary strategy is used for the second half of the backtrack limit. This article presents a qualitative ATPG cost model based on the number of test generation events, uses this model to explain why search strategy switching is faster, and shows experimental evidence to verify both the cost model and search strategy switching theory. The experiments were performed with the ISCAS circuits and our implementation of the FAN algorithm.