Theoretical aspects of importance sampling applied to false alarms

A study is made of the theoretical improvements which are possible using importance sampling techniques to estimate iow false alarm rates in a radar simulator. Conditions are shown where significant improvements can be obtained. In realistic situations importance sampling can dramatically decrease the number of Monte Carlo runs from impractical orders of magnitude such as
10^{10}
to practical numbers of the order of 1000.