Environmetrics, meteorology, seismology, and geo-statistics are frequently cited areas where the efficiency of observing/ monitoring networks is crucial, both from statistical and economical points of view. Monitoring of large computer networks is one of the most recent and notable examples, in which the efficient choice of sites to be monitored is a crucial component of empirical studies related to performance evaluation. The presented approach is based on the expansion of covariance kernels with respect to their eigenfunctions and consequent use of results developed in convex design theory.
展开▼
