How this discretization scheme converges in probability to the stochastic integral $\int_0^T H_t \mathrm d B_t$?

33 Views Asked by At

Let $(\Omega, \mathcal F, \mathbb P)$ be a probability space. Let $(H_t, t \in [0, T])$ be an adapted process and $(B_t, t \ge 0)$ the standard Brownian motion w.r.t. the same filtration $(\mathcal F_t, t \ge 0)$. For $n \ge 1$, we define $(H^{(n)}_{t}, t \in [0, T])$ by $$ H^{(n)}_{t} := \sum_{i=1}^{2^n} H\left(\frac{(i-1) T}{2^n}\right) 1_{ \left (\frac{(i-1) T}{2^n}, \frac{i T}{2^n} \right]}(t). $$

So $H^{(n)}_{t}$ is a step function of time. We define $$ (H^{(n)} \cdot B)_T := \sum_{i=1}^{2^n} H\left(\frac{(i-1) T}{2^n}\right) \left ( B\left(\frac{i T}{2^n}\right) - B\left(\frac{(i-1) T}{2^n}\right) \right ) \quad \forall n \ge 1. $$

For ease of notation, let $t_i := iT/2^n$. Then $$ (H^{(n)} \cdot B)_T = \sum_{i=1}^{2^n} H_{t_{i-1}} (B_{t_i} - B_{t_{i-1}}). $$

It is mentioned (without proof) at page 37 of this note that

Theorem If $(H_t, t \in [0, T])$ has continuous trajectories and $\mathbb E[ \int_0^T H_t^2 \mathrm d t] < \infty$, then $$ (H^{(n)} \cdot B)_T \underset{n \rightarrow \infty}{\longrightarrow} (H \cdot B)_T \quad \text{in probability}, $$ where $(H \cdot B)_T \equiv \int_0^T H_t \mathrm d B_t$ is the Itô's stochastic integral.

This convergence result is then used in a version of Itô's lemma at page 39 of the same note.

My question Could you explain how to prove this theorem?

1

There are 1 best solutions below

0
On BEST ANSWER

This is Proposition 5.9 in Le Gall's Brownian Motion, Martingales, and Stochastic Calculus.


Proposition 5.9 Let $X$ be a continuous semi-martingale, and let $H$ be an adapted process with continuous sample paths. Then, for every $t>0$, for every sequence $0=t_0^n<\cdots<t_{p_n}^n=t$ of subdivisions of $[0, t]$ whose mesh tends to $0$, we have $$ \lim _{n \rightarrow \infty} \sum_{i=0}^{p_n-1} H_{t_i^n}\left(X_{t_{i+1}^n}-X_{t_i^n}\right)=\int_0^t H_s \mathrm{~d} X_s, $$ in probability.