3
$\begingroup$

does someone know how to solve the following SDE

$$dS_t=(\alpha S_t+f(t))dW_t, S_0=s$$

where $f(t)$ is a deterministic function and $W_t$ is a standard brownian motion. Is there a explicit solution for this SDE? Many thanks!

$\endgroup$
3
$\begingroup$

Note that the SDE $$dS_t = (\alpha \cdot S_t+f(t)) \, dW_t \qquad S_0 = s$$ is a linear SDE and therefore there is indeed an explicit formula for the solution of the SDE. One approach is the following:

  1. Solve the homogeneous SDE $$dS_t = \alpha \cdot S_t \qquad X_0 = x$$ There are several possibilities to do so: Either you have some intuition how the solution might look like or you apply Itô's formula to $X_t := g(W_t)$ to obtain conditions on (the derivatives of) $g$ allowing you to determine $g$.
  2. Solve the inhomogeneous SDE $$dS_t = (\alpha \cdot S_t+f(t)) \, dW_t$$ One approach is the so-called variation of constants: Let $S_t^0$ such that $\frac{1}{S_t^0}$ solves the homogeneoues SDE, $S_0^0 = 1$ (we solved this SDE in the first step, so there is an explicit formula). Now we apply Itô's formula to $$Z_t := S_t \cdot S_t^0$$ which works fine since we know $dS_t$, $dS_t^0$. Therefore, we can determine $Z_t$ and consequently also $S_t = \frac{Z_t}{S_t^0}$.

Literature: e.g. René L. Schilling/Lothar Partzsch: Brownian Motion - An Introduction to Stochastic Processes

$\endgroup$
0
$\begingroup$

It is easy to check that \begin{align*} d\left(e^{-\frac{\alpha^2}{2}t - \alpha W_t } S_t \right) = e^{-\frac{\alpha^2}{2}t - \alpha W_t } f(t) dW_t. \end{align*} Then, $S_t$ can be solved subsequently.

$\endgroup$
  • $\begingroup$ What is a motivation behind considering the discounting factor? $\endgroup$ – Idonknow Dec 11 '19 at 15:39
  • $\begingroup$ It is the integration factor, and is motivated by the usual treatment for mean-reverting processes. $\endgroup$ – Gordon Dec 11 '19 at 16:24
  • $\begingroup$ Thanks for your reply. Is the integrating factor same as in first order ordinary differential equation? $\endgroup$ – Idonknow Dec 11 '19 at 23:39
  • $\begingroup$ Yes. Very similar idea. $\endgroup$ – Gordon Dec 12 '19 at 12:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.