Explicit stochastic integral representation of path-dependent Brownian functionals (Q6985683)

Let \(B=(B_t),t\in [0,T]\) be a Brownian motion defined on a probability space \((\Omega,\mathfrak{J},P)\). For any non-negative integer \(n\), define the following path-dependent Brownian functional: \N\[\NF(n):=\left(\int_0^TB_sds\right)^{2n+1}. \N\]\NThe functionals \([F(n)]^+,[F(n)]^-\) and \(|F(n)|\) are not stochastically (in Malliavin sense) smooth, and both the well-known Clark-Ocone formula [\textit{D. Ocone}, Stochastics 12, 161--185 (1984; Zbl 0542.60055)] and its generalization, the Glonti-Purtukhia formula [\textit{O. A. Glonti} and \textit{O. G. Purtukhia}, Theory Probab. Appl. 61, No. 1, 133--139 (2017; Zbl 1386.60283); translation from Teor. Veroyatn. Primen. 61, No. 1, 158--164 (2016)], are inapplicable to them. In this paper, the authors use a certain modification of the earlier generalization of the Clark-Ocone formula by \textit{V. Jaoshvili} and \textit{O. Purtukhia} [``Stochastic integral representation of functionals of Wiener processes'', Bull. Georg. Natl. Acad. Sci. 171, No. 1, 17--20 (2005)] and derive a stochastic integral representation with an explicit form of the integrands.