Limit theorems for random walk under the assumption of maxima large deviation (Q2882293)

The main purpose of this paper is to establish conditioned limit theorems for different statistics (under condition \(M_n\geq \theta n\)) of a random walk with steps obeying the Cramer condition and to prove the functional limit theorem for such a random walk; here \(M_n=\max_{k\leq n}S_k,\;S_n=\sum_1^n X_l,\;n=1,2,\dots,\;S_0=0,\;X_l-\) i.i.d.r.v.'s. It was showed also that \(n-\tau_n\;(\tau_n=\min\{k:\;S_k=M_n\})\) conditionally converges to an exponential r.v. in distribution when \(M_n\geq \theta n,\) so the limit process in the functional limit theorem for the processes set by \(S_n\) in the case \(M_n\geq \theta n\) is the same as in the case \(S_n\geq \theta n,\) namely, the Brownian bridge. To motivate the abovementioned results the author used two equivalences, established for \(P(S_n\geq \theta n-u)\) in \textit{V. V. Petrov} [Teor. Verojatn. Primen. 10, 310--322 (1965; Zbl 0235.60028); translation in Theor. Probab. Appl. 10, 287--298 (1965)] and for \(P(M_n\geq \theta n-u)\) in 2001 by \textit{A. A. Borovkov} and \textit{D. A. Korshunov} [Theory Probab. Appl. 46, No. 4, 603--618 (2001); translation from Teor. Veroyatn. Primen. 46, No. 4, 640--657 (2001; Zbl 1036.60021)].