Riemann-Hilbert problem (source code)

= Riemann-Hilbert problem
{c}
{title2=$F_+-F_-=J$}

A <Riemann-Hilbert problem> seeks <holomorphic functions> off an oriented <contour>, with prescribed relations between their limiting <boundary values> and with specified behavior at infinity, endpoints or other distinguished points. The plus side lies to the left when following the contour orientation; the minus side lies to the right.

A standard matrix form prescribes a <jump matrix> $J$ and asks for
$$
M_+(z)=M_-(z)J(z),\qquad z\in\Gamma,\qquad M(z)\to I\quad(z\to\infty).
$$
The <matrix> multiplication order is part of the problem. Some versions also prescribe poles and their residues. The contour, jump, normalization and local singularity conditions together determine the analytic problem.<Riemann-Hilbert Problems by Percy Deift>[Lecture 1, pp. 5–6, contour sides and matrix jump definition]{ref}

For an <additive Riemann-Hilbert problem>, the condition is $F_+-F_-=g$. On a positively oriented smooth closed contour, the <Cauchy integral>
$$
F(z)=\frac1{2\pi i}\int_\Gamma\frac{g(\zeta)}{\zeta-z}\,d\zeta
$$
has that jump under the usual regularity assumptions, by the <Sokhotski–Plemelj theorem>. For the <unit circle> and $g=1$, it gives $F=1$ inside and $F=0$ outside. This checks the sign convention and the vanishing exterior normalization. A multiplicative matrix jump generally needs additional analysis beyond this scalar Cauchy formula.