If the Ito process $\mathrm{d}X_{t}=\mu(X_{t})\mathrm{d}t+\sigma (X_{t})\mathrm{d}W\equiv0$, it seems that $\mu(X_{t})=0$ and $\sigma(X_{t}% )=0$. Is it true? If yes, how to prove? We know that $\mathrm{d}X_{t}% =\mu(X_{t})\mathrm{d}t+\sigma(X_{t})\mathrm{d}W\equiv0$ exactly means given time $s$ with $X_{s}\neq0$ $$ \forall u>s\qquad0\equiv X_{u}-X_{s}=\int_{s}^{u}\mu(X_{t})\,\mathrm{d}% t+\int_{s}^{u}\sigma(X_{t})\,\mathrm{d}W_{t}% $$ then $\mu(X_{t})\equiv0$ and $\sigma(X_{t})\equiv0$ for $s<t<u$.
2026-04-04 08:34:39.1775291679
If the SDE of an Ito process is identically zero, are the drift and volatility both identically zero?
155 Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail At
2
It is true whether $X_{s}=0$ or not
Remark: If $\int_{s}^{x}f(t)\,\mathrm{d}t=0$ for any $x>s$, then $f(x)=\frac{\mathrm{d}}{\mathrm{d}x}\left( \int_{s}^{x}f(t)\,\mathrm{d}% t\right) =\frac{\mathrm{d}}{\mathrm{d}x}(0)=0$
For $X_{t}=X_{s}$ is constant for any $t>s$, the quadratic variation is zero $$ 0=[X_{t}]=\int_{s}^{u}\sigma^{2}(X_{t})\,\mathrm{d}t $$ which means $\sigma(X_{t})=0$. And for any $u>s$ $$ \int_{s}^{u}\mu(X_{t})\,\mathrm{d}t=(X_{u}-X_{s})-\int_{s}^{u}\sigma (X_{t})\,\mathrm{d}W_{t}=0-0=0 $$ thus $\mu(X_{t})=0$.