/-
Copyright 2026 The Formal Conjectures Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
https://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-/modulepublicimportFormalConjecturesUtil
Equivalents to the Riemann Hypothesis
This file contains formal statements that are known to be equivalent to the Riemann Hypothesis,
following the survey [Co26], §4. The statement of the Riemann Hypothesis itself is in
FormalConjectures.Millennium.RiemannHypothesis.
Robin's Criterion. The Riemann Hypothesis is equivalent to
$$
\sigma(n) < e ^ {\gamma} n\log\log n
$$
for all $n > 5040$, where $\gamma$ is the Euler-Mascheroni constant and $\sigma(n)$ is
the sum of divisors.
Lagarias' Criterion. A refinement of Robin's criterion: the Riemann Hypothesis
is equivalent to
$$
\sigma(n) < H_n + e^{H_n}\log H_n
$$
for all $n \geq 2$, where $H_n = 1 + 1/2 + \cdots + 1/n$ is the $n$th harmonic number.
Note that [Co26] states this for all $n \geq 1$, but at $n = 1$ both sides equal $1$; Lagarias'
original formulation is the non-strict inequality for all $n \geq 1$, with equality only at
$n = 1$.
The nontrivial zeros of the Riemann zeta function with positive imaginary part. Since the
nontrivial zeros are closed under complex conjugation, these determine all nontrivial zeros.
The numbers
$$
\lambda_n := \sum_{\rho}\left(1 - \left(1 - \frac{1}{\rho}\right)^n\right)
$$
appearing in Li's criterion, summing over the nontrivial zeros $\rho$ of the
Riemann zeta function.
This is a real number because nontrivial zeros always come in conjugate pairs.
Note that, as stated, the sum is only conditionally convergent, however it is absolutely convergent
when taken in conjugate pairs, since $1/\rho + 1/\bar\rho = 2\Re\rho/|\rho|^2$. Hence we sum
over the zeros with positive imaginary part with summand 2 * (_).re, which makes the unordered
sum ∑' meaningful.
Li's Criterion. The Riemann Hypothesis is equivalent to the positivity of
$$
\lambda_n := \sum_{\rho}\left(1 - \left(1 - \frac{1}{\rho}\right)^n\right)
$$
for each $n > 0$.
Beurling-Nyman Criterion. The Riemann Hypothesis is equivalent to the statement that
the set of finite linear combination $\sum_{\nu = 1}^n c_{\nu}\rho_{\theta_{\nu}}$ is dense
in $L^2(0, 1)$, when ranging over $0 < \theta_{\nu}\leq 1$ and
$\sum_{\nu = 1}^n c_{\nu}\theta_{\nu} = 0$.
The additive condition on test functions. The property in [Co26] that $\hat{g}(\pm i/2) = 0$,
where $\hat{g}$ is the Mellin transform, corresponds to the vanishing of the Fourier transform
at $\pm 1/2$.
Weil's Positivity Criterion. The Riemann Hypothesis is equivalent to the non-positivity of
Weil factor sums on the convolution of test functions. In [Co26] this is expressed as
$$
\sum_v W_v(g\star g^) \leq 0
$$
for all compactly-supported smooth functions $g : \mathbb{R}_+^{\times} \to \mathbb{C}$ such that
the Mellin transform of $g$ vanishes at $\pm i / 2$, where $g^(x) := \bar{g}(x^{-1})$ and $W_v$
are the multiplicative Weil factors defined in [Co26]. We formalise this as the corresponding
additive statement (via $x = e^u$), since convolution in mathlib is additive, and restrict to
real-valued test functions $g$, which suffices for the equivalence.