5
$\begingroup$

Consider the equation

$$f'(x)+ g(x)f(x)=0$$

This equation is an ODE and has a solution $$ f(x)=C e^{ \int_1^x g(x) \ dx}.$$

Similarly, we can look at complex variables and consider the equation and Wirtinger derivatives

$$ (\partial_{\bar z} f)(z) +g(z) f(z)=0.$$

Can one still write down an explicit solution?

$\endgroup$
2
  • $\begingroup$ Is $f(z)$ analytic? If so then the derivative of $f(z)$ with respect to $\bar{z}$ does not exist unless $f(z)$ is a constant. $\endgroup$
    – asymptotic
    Sep 23, 2020 at 13:59
  • $\begingroup$ No, it is not analytic... $\endgroup$
    – Sascha
    Sep 23, 2020 at 14:06

1 Answer 1

6
$\begingroup$

You can start by looking at the chain rule for wirtinger derivatives, from which you deduce that

$$ \partial_{\bar z} \exp(h(z)) = \exp(h(z)) \cdot \partial_{\bar z} h(z) $$

Therefore, if you find a function $h$ such that $\partial_{\bar z} h = - g(z)$ (I think you forgot a "$-$" sign in your solution for the real case!) taking $f(z) = \exp(h(z)) $ will solve your problem. In general, this is known as the d-bar problem (or $\bar\partial-$problem). As Daniele points out this Q&A is a god resource for the $\bar\partial-$problem in 1 dimension.

$\endgroup$
2
  • 3
    $\begingroup$ @Sascha, Jaume for a solution of the $1$-dim $\bar\partial$-problem this Q&A may be relevant. $\endgroup$ Sep 23, 2020 at 14:49
  • 1
    $\begingroup$ Indeed @daniele-tampieri! That's a way better link to the $\bar \partial$ problem than the wikipedia page that has (surprisingly?) almost no reference $\endgroup$
    – Jaume
    Sep 23, 2020 at 15:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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