Asymptotic behavior of the quadratic variation for trajectories of processes with independent increments (Q801610)

Let \(\{X_ n\}^{\infty}_{n=1}\) be a sequence of independent random variables with expectations \(EX_ i=0\), \(i=1,2,..\).. Set \(S_ n=\sum^{n}_{j=1}X_ j\) and \(T_ n=\sum^{K(n)}_{p=1}(S_{i_{p}(n)}-S_{i_{p-1}(n)})^ 2\) where \(i_ 0(n)=0<i_ 1(n)<...<i_{k(n)}(n)=n\) is a sequence of integers depending on n. Under the conditions \(C_ 1a(n)\leq i_ p(n)-i_{p- 1}(n)\leq C_ 2a(n)\); \(a(\tau)/a(t)\leq \phi (\tau t^{-1})\), t,\(\tau\in (0,\infty)\); \(EX^ 2_ i>C_ 3\); \(E| X_ i|^{2+\delta}<C_ 4\) where \(C_ 1,C_ 2,C_ 3,C_ 4\), \(0<\delta <1\) are constants, \(a(n)=n/k(n)\) and \(\phi\) (t) is a nondecreasing function on (0,\(\infty)\), the author proves that there exists a constant A \((0<A<\infty)\) such that \(\limsup_{n\to \infty}T_ n/(n+a(n)\ell n\ell n n)=A\) with probability 1.