Subharmonic function (source code)

= Subharmonic function
{wiki}

= Subharmonic
{synonym}

For $u\in C^2(\Omega)$, <subharmonic> means $\Delta u\geq0$, using the <Laplacian> $\Delta=\sum_j\partial_{x_j}^2$. The strict inequality $\Delta u>0$ defines strict subharmonicity and is stronger. Every <harmonic function> is therefore <subharmonic>, even though its <Laplacian> vanishes.<Partial Differential Equations by John K. Hunter>[chapter2, section2.1.1, Definition2.4 and Theorem2.5, printed p22]{ref}

The general definition allows an upper-semicontinuous <function> $u:\Omega\to[-\infty,\infty)$, not identically $-\infty$ on any connected component, satisfying the ball mean inequality
$$
u(x)\leq\frac1{|B_r|}\int_{B_r(x)}u(y)\,dy
\qquad(\overline{B_r(x)}\subset\Omega).
$$
For locally integrable <functions>, distributional nonnegativity of $\Delta u$ produces this canonical representative. An arbitrary change at one point need not preserve the mean inequality, which is why the representative matters.

The example $u(x)=|x|^2$ has $\Delta u=2n>0$, while $|z|^p$ is <subharmonic> for every $p>0$. More generally, $|f(z)|^p$ is <subharmonic> for a <holomorphic function> $f$ and $p>0$. At a strict interior maximum of a twice differentiable <function>, the <Hessian matrix> is negative semidefinite; the <maximum principle for subharmonic functions> extends this elementary obstruction to the non-strict and nonsmooth setting.