Abstract

This paper addresses the problem of ranking and selection for stochastic processes,su h as target tracking algorithms, where variance is the performance metric. Comparison of di#erent tracking algorithms or parameter sets within one algorithm relies on time-consu3J9 andcompu9yj3J3J9y demanding simu lations. We present a method to minimize simuy3F90 time, yet to achieve a desirable confidence of the obtainedresune byapplying ordinal optimization and compu07y buu0 allocation ideas and techniqu80 while taking intoaccou t statistical properties of the variance. The developed method is applied to a general tracking problem of N s sensors tracking T sing a sequ# tial mu lti-sensor datafuayD tracking algorithm. The optimization consists of finding the order of processing sensor information thatresuDF in the smallest variance of the position error. Resur. that we obtained with high confidence levels and inredu99 simu lation times confirm the findings from ou previou research (where we considered onlytwo sensors) that processing the best available sensor the last performs the best, on average. The presented method can be applied to anyranking and selection problem where variance is the performance metric.

Citations