GENETIC ALGORITHM FOR TEST SCHEDULING WITH DIFFERENT OBJECTIVES

Test scheduling is one of the hard optimization problems involved in a ny VLSI design process. Depending on the overall design objectives, th e objective of test scheduling may differ. Since it is computationally infeasible to solve this problem optimally, different approximation s chemes are employed for different objectives resulting in different fo rmulations. In this paper we show that test scheduling with different objective functions can be formulated and approximately solved in a co mmon framework using genetic algorithms. The genetic algorithm is pres ented and experimental results for three different formulations are sh own. The results show that the proposed genetic algorithm provides a u niform framework for obtaining near optimal solutions for the test sch eduling problem with different objectives. Comparison with branch and bound method solutions reveal the superiority of the genetic algorithm .