Riemann Hypothesis

Is there a hidden regularity in the distribution of prime numbers, and why does uncovering it require going into the complex plane and counting zeros of a holomorphic function defined by an infinite series?

• **Millennium Prize Problem:** One of 7 Clay Institute problems with a $1,000,000 prize. In 2023, Platt and Trudgian verified GRH for the first 3*10^12 zeros.
• **Cryptography:** GRH gives the best bounds on factoring algorithm complexity (GNFS). Miller-Rabin is deterministic under GRH.
• **Quantum chaos:** Zero statistics of zeta(s) match GUE random matrix statistics - the same laws as quantum energy levels of chaotic systems.
• **Primality testing:** AKS runs in O(log^{7.5} n); deterministic Miller-Rabin in O(log^2 n) under GRH.

Zeta Function and the Euler Product

The Riemann zeta function zeta(s) was defined by Euler for real s > 1 as the sum of n^{-s}. Riemann in 1859 showed it admits analytic continuation to the whole complex plane (except a pole at s=1) and derived a functional equation. The Riemann Hypothesis - that all non-trivial zeros of zeta(s) lie on the line Re(s) = 1/2 - remains unproved after 165+ years and connects the behavior of primes to the geometry of the complex plane.

Connection to quantum mechanics: Montgomery (1973) showed the pair correlation of zeta zeros matches the statistics of GUE (Gaussian Unitary Ensemble) random matrices. The Hilbert-Polya conjecture: if a self-adjoint operator with zeta zeros as its spectrum exists, GRH follows from the reality of the spectrum. This connects zeta(s) to quantum chaos.

Practical consequences of GRH: (1) The smallest primitive root mod p is O(log^6 p). (2) Miller-Rabin is deterministic under GRH in O(log^2 n). (3) Equivalent formulation: |pi(x) - Li(x)| < (1/(8*pi)) sqrt(x) log(x) for x >= 2657. The AKS primality test runs in O(log^{7.5} n).

The trivial zeros s = -2, -4, -6, ... are not the subject of GRH. The result Re(rho) < 1 (Hadamard, 1896) is distinct from GRH. GRH asserts Re(rho) = 1/2 exactly. Partial results: Hardy (1914) proved infinitely many zeros lie on the critical line; Conrey (2000) proved more than 40 percent of zeros are on Re(s) = 1/2.

Structure of Zeros and Prime Distribution

The zeros of the Riemann zeta function act as 'frequencies' in a Fourier-like decomposition of the prime-counting function. Each zero rho contributes a term x^rho/rho to the explicit formula for psi(x). If all zeros lie on Re(s) = 1/2, these contributions oscillate rather than grow, giving the optimal error O(x^{1/2} log^2 x). A zero off the critical line would cause larger oscillations in pi(x) and psi(x).

Alternative approaches to GRH: spectral theory (Hilbert-Polya), motivic cohomology, function-field analogues (Weil, 1948, proved for curves over finite fields via intersection theory). The Weil conjectures (proved by Deligne, 1974) are the geometric analogue. Despite these analogues, no approach has yet succeeded for the classical zeta function.

Connections to other areas of mathematics

The Riemann Hypothesis is central to modern mathematics, intersecting number theory, random matrices, and spectral theory.

• Prime number theory — Related topic
• Random matrices — Related topic
• Dirichlet L-functions — Related topic
• Algebraic geometry — Related topic

Итоги

• zeta(s) = sum n^{-s} = prod (1-p^{-s})^{-1} for Re(s) > 1; the Euler product connects complex analysis to prime distribution
• Functional equation: zeta(s) = 2^s pi^{s-1} sin(pi*s/2) Gamma(1-s) zeta(1-s) - symmetry about Re(s) = 1/2
• Trivial zeros: s = -2, -4, -6, ...; non-trivial zeros lie in the critical strip 0 < Re(s) < 1
• GRH: all non-trivial zeros have Re(s) = 1/2 (the critical line); verified for the first 3*10^12 zeros (2023), unproved in general
• Under GRH: psi(x) = x + O(x^{1/2} log^2 x) - optimal error bound in the prime number theorem