= Solution
Set $F_t=e^{\alpha B_t-\alpha^2t/2}$, a positive <stochastic exponential> with $dF_t=\alpha F_t\,dB_t$. The <Itô formula> gives
$$
d(F_t^{-1})=-\alpha F_t^{-1}dB_t+\alpha^2F_t^{-1}dt.
$$
For $Y=X/F$, the <Itô product rule> and $d[X,F^{-1}]_t=-\alpha^2X_tF_t^{-1}dt$ cancel both the Brownian terms and the extra drift. Consequently
$$
\boxed{dY_t=F_t^{\delta-1}Y_t^\delta\,dt,\qquad Y_0=x_0.}
$$
This is an ordinary <Bernoulli differential equation> with continuous random coefficients. For $\delta\ne1$, separate variables to obtain
$$
Y_t^{1-\delta}=x_0^{1-\delta}+(1-\delta)\int_0^tF_s^{\delta-1}ds.
$$
Therefore the <Bernoulli stochastic differential equation> has the explicit maximal positive solution
$$
\boxed{X_t=F_t\left[x_0^{1-\delta}+(1-\delta)\int_0^tF_s^{\delta-1}ds\right]^{1/(1-\delta)},\quad\delta\ne1,}
$$
valid while its bracket remains strictly positive. When $\delta=1$, the reduced equation is $dY_t=Y_tdt$, and
$$
\boxed{X_t=x_0\exp\!\left(\alpha B_t+\left(1-\frac{\alpha^2}{2}\right)t\right).}
$$
Substituting these expressions back through the product rule verifies the original <stochastic differential equation>, not just the reduced equation.
Back to article page