Prove that process is uniformly integrable

1.1k Views Asked by At

Let $(X_t)_{t\ge 0}$ be a stochastic process, and let $Y$ be an integrable random variable, such that $|X_t|\le Y$ for $t\ge0$. Prove that $(X_t)_{t\ge 0}$ is uniformly integrable.

From definition, we have that $(X_t)_{t\ge 0}$ is uniformly integrable if $$\sup_{t\in [0,\infty)}\int_{\{X_t > \epsilon \}} |X_t|d\mathbb{P} \rightarrow 0$$ when $\epsilon\rightarrow \infty$.

I think I can use here Dominated convergence theorem.

Define a sequence of functions: $Y_t^{\epsilon}:=X_t \mathbb{1}_{\{X_t > \epsilon \}}$

Because $|X_t|$ is bounded by $Y$ and $Y$ is integrable, we have that $(Y_t^{\epsilon})_{\epsilon} \rightarrow 0$

From Dominated convergence theorem we have that $\int_{\{X_t > \epsilon \}} |X_t|d\mathbb{P} \rightarrow 0$ as $\epsilon \rightarrow \infty$

I'm very doubtful about my solutions, itts probably wrong. Can you explain why and what I should do instead?

1

There are 1 best solutions below

1
On BEST ANSWER

The problem is that you showed the result for a fixed $t$, but not that it holds uniformly in $t$.

Hint: note that for each $t$ and $\varepsilon$, the inequality $$|X_t|\mathbb 1_{\{ |X_t|\gt\varepsilon\}} \leqslant Y\mathbb 1_{\{ Y\gt\varepsilon\}} $$ holds. Integrate and conclude by monotone convergence.