Algorithms for cost optimised test strategy selection

This paper describes a system which aids designers in the selection of economically optimal design for testability strategies. The approach recognizes that design for testability decisions taken for parts of the circuit affect subsequent decisions for the design as a whole. Economics models are used to objectively evaluate strategies, and a set of algorithms is presented which provide efficient methods for searching the test strategy space. Different approaches are compared and a novel use of simulated annealing for test strategy planning is presented. Examples on real designs are given, with an indication of CPU time and computational complexity