An optimal software debugging model

This paper deals with the question of optimal debugging times for large-scale computer programs, given certain assumptions concerning the distribution of error within the program, the cost that is incurred whenever a fault occurs, and the cost of debugging. Modeling takes place within the framework of independent Poisson processes describing faults caused by errors. The optimization aim is to minimize the long-run total expected discounted cost. Some general results concerning the optimal polity are presented.This research has been supported by NATO Grant No. RG.86/0024, by the Alexander von Humboldt Foundation, and by the University of Passau, Germany.