2
$\begingroup$

I'm reading about the Gauss hypergeometric differential equation: \[ x(1-x)y'' + (c - (a + b + 1)x)y' - aby =0 \]

Apparently this is a kind of Fuchsian differential equation: \[ \left(\begin{array}{c} \frac{dy}{dx} \\\\ \frac{dz}{dx} \end{array} \right) = \left(\begin{array}{cc} 0 & 1 \\\\ -ab & \frac{c_1}{x} + \frac{c_2}{x-1}\end{array} \right) \left(\begin{array}{c} y \\\\z \end{array} \right) \]

whose sections define a local system over a sphere with 3 punctures, $\mathbb{C}\backslash \{0,1 \infty \}$. We can worry about the monodromy of ODE solutions about these three points $0,1,\infty$.

The paper I'm reading says Fuchsian equations may be is solved by a Selberg-like integral.

\[ y(x)= \int_\Delta s^{c-b}(s-1)^{c-a-1} (s-x)^{-b} \, ds \]

Where the integral is over a "twisted cycle" $\Delta$ in $H_1(\mathbb{C}\backslash\{ 0,1,x\}, L^* )$.


My trouble is understanding what these twisted cycles look like. The local system $L^*$ is dual to the local system $L$ generated by sections $s \mapsto s^{c-b}(s-1)^{c-a-1} (s-x)^{-b} $, which is multivalued. The values form a discrete set with (possibly countably) many fibers over a point in $\mathbb{C}\backslash\{ 0,1,x\}$ so $L$ should be dual to a branched cover over the three-punctured sphere possibly with more information. Then maybe $H_1(\mathbb{C}\backslash\{ 0,1,x\}, L^*)$ is the collection of loops in this branched cover.


How is $H_1(\mathbb{C}\backslash\{ 0,1,x\}, L^*)$ different from $H_1(\mathbb{C}\backslash\{ 0,1,x\})$ ?

$\endgroup$

0

Your Answer

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

Browse other questions tagged or ask your own question.