On an estimate of probability of exceeding a line by weighted aggregate of sub-Gaussian random process (Q6600527)

The authors deals with the problem of estimating the probability of exceeding a level given by a line \(ct\),\(\ c>0\), by trajectories of a sum of sub-Gaussian random processes \(X_i(t)\), \(i=\overline{1,n}\), defined on a compact set \(B\) with certain weighting functions \(w_i(t)\). They derived the upper estimates of the probabilities \({\mathbf{P}}\left\{{\mathop{{\sup}}_{t\in B} \left(\sum^n_{i=1}{w_i\left(t\right)X_i(t)}{-}ct\right)\ }>x\right\}\), \({\mathbf{P}}\left\{{\mathop{{\inf}}_{t\in B} \left(\sum^n_{i=1}{w_i\left(t\right)X_i(t)}{-}ct\right)\ }<-x\right\}\), \({\mathbf{P}}\left\{{\mathop{{\sup}}_{t\in B} \left\vert \sum^n_{i=1}{w_i\left(t\right)X_i(t)}{-}ct\right\vert \ }>x\right\}\). Using the method of metric entropy, the results obtained by \textit{O. Vasylyk} et al., Random Oper. Stoch. Equ. 13, No. 2, 111--128 (2005; Zbl 1118.60025), for a class of \(\Phi\)-sub-Gaussian random processes are generalized and improved. As an example, the derived estimate is applied to the average sum of sub-Gaussian Wiener random processes, i.e. random processes that have the same covariance function as the (Gaussian) Wiener process, but with sub-Gaussian trajectories. This obtained estimates can be applied directly in the queuing theory in estimating the finite size \(x>0\) buffer overflow probability with linear service intensity, as well as in insurance mathematics in estimating the bankruptcy probability for the corresponding risk process.