The use of statistical experimental design techniques for the systematic improvement of an automated, heuristic targeting algorithm

Abstract

Force application, or rapid targeting or retargeting of weapons, is necessary to keep at risk high value military movable targets (MT) with a minimum number of weapons. Because of the targets' mobility, the system should be automated and able to deal with rapidly changing scenarios. If a multiple-warhead weapon is used, the targets must be arranged into sets of targets, each set being targeted by one weapon. The number of target sets formed must be kept to a minimum. This paper presents a heuristic, automated minimization algorithm created for that purpose. Statistical experimental design and optimization techniques were used to improve the efficiency of the algorithm in a wide variety of test scenarios by systematically selecting the best combination of algorithmic variations. Using a randomized block factorial design and numerous test scenarios with varying target distributions, competing versions of the algorithm were compared and the best combination of rules produced. The results indicate that experimental design techniques can be applied to heuristic rules to improve them in a systematic and unbiased way.