Some properties of the generalized TTT transform (Q1015865)

For a distribution function \(F\), and a positive function \(\varphi\) on \((0,1)\), the generalized total time on test (GTTT) transform is defined as \(H_F^{-1}(t;\varphi)=\int_0^{F^{-1}(t)}\varphi(F(x))\,dx\) for \(t\in(0,1)\). The inverse of the GTTT transform is a distribution function on \((0,1)\). The authors show that the stochastic dispersive order and the stochastic convex transform order are invariant under the GTTT transforms in the sense that if the distribution functions \(F\) and \(G\) are such ordered, then the corresponding inverses of the GTTT transforms, \(H_F(t;\varphi)\) and \(H_G(t;\varphi)\), are also such ordered. The authors also derive a result that involves two GTTT transforms that are based on the same distribution function \(F\), but on two different positive functions \(\varphi_1\) and \(\varphi_2\).