Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Given the following Lemma: Let $A_{t}=\int_{0}^{t}a_{s}dB_{s}$ where $a$ is an adapted process satisfying $\mathbb{P}\Big(\int_{0}^{T}a^{2}_{u}du < \infty\Big) = 1$ and $B$ is a standard Brownian motion. $A$ is a supermartingale if $A_{t}(\omega)$ is bounded by a constant $K$ for all $t,\omega$.

For a given adapted process $\eta$ consider the following stochastic differential equation: $$dZ_{t}:=-Z_{t}\eta_{t}dB_{t},\;\;\;Z_{0}:=1.$$ -Explain why we have $$(Z_{t})^{\beta}=e^{-\beta\int_{0}^{t}\eta_{u}dB_{u}-\frac{1}{2}\beta\int_{0}^{t}\eta^{2}_{u}du}$$ where $\beta$ is a positive constant. (I've gotten this one.)

-Show that $$\mathbb{E}\Bigg[\Big(e^{-\beta\int_{0}^{t}\eta_{s}dB_{s}-\beta^{2}\int_{0}^{t}\eta^{2}_{s}ds}\Big)^{2}\Bigg ]\leq 1$$ where $\beta$ is a positive constant and $\eta$ is an adapted process, using the given lemma. (The given lemma only seems to apply to part of the exponent?)

-Use the Cauchy-Schwartz inequality to derive the expression $$\mathbb{E}[Z_{t}^{\beta}]\leq \mathbb{E}\Bigg[e^{\beta(2\beta - 1)\int_{0}^{t}\eta_{u}^{2}du}\Bigg]^{\frac{1}{2}}$$ where $\beta$ is a positive constant. (Where did the $\beta(2\beta-1)$ come from?)

-Use the previous question with $\beta := 2$, to show that for $\eta$ bounded, we have for all $T>0$ $$\mathbb{E}\Bigg[\int_{0}^{T}(Z_{t}\eta_{t})^{2}dt\Bigg]<\infty,$$ and conclude that for a bounded $\eta$, the process $Z$ is a martingale. (Totally confused)

-Define the process $\eta$ by $\eta_{t}:=\frac{1}{R_{t}}$ where $R$ is the Bessel process, defined as the solution to the following equation$$dR_{t}:=\frac{1}{R_{t}}dt+dB_{t},\;\;\;R_{0}:=1.$$ Compute the differential $d\Big(\frac{\eta_{t}}{Z_{t}}\Big)$ and conclude that $Z_{t}=\eta_{t}$. (I've worked on this but didn't get the right conclusion.)

share|improve this question
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.