Typical short-interval bounds for the maximum of the Riemann zeta function

From papers

Let ζ(s)\zeta(s) denote the Riemann zeta function. For T>0T>0, consider its maximum on the interval [T,T+2π][T,T+2\pi].

Typical short-interval maximum conjecture. Let ϵ>0\epsilon>0. For T1T_1 sufficiently large depending on ϵ\epsilon, for all T1T2T1T_1\leq T\leq 2T_1, outside a set of “bad” values of TT of measure at most ϵT1\epsilon T_1, one has

exp(loglogT(2+o(1))logloglogT)maxTtT+2πζ(1/2+it)exp(loglogT(1/4+o(1))logloglogT).\exp(\log\log T-(2+o(1))\log\log\log T) \leq \max_{T\leq t\leq T+2\pi}|\zeta(1/2+it)| \leq \exp(\log\log T-(1/4+o(1))\log\log\log T).

This is motivated by comparing the prime-sum model for logζ(1/2+it)\log|\zeta(1/2+it)| with a log-correlated random model on a typical short interval. The assertion gives separate lower and upper bounds rather than the sharper conjectural asymptotic above, and its status is not resolved here.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Primary source

Adam J. Harper, “A note on the maximum of the Riemann zeta function, and log-correlated random variables”, arXiv:1304.0677 (2013).

Solutions 0

No solutions have been posted yet.