Discrete Schwartz distributions and the Riemann zeta-function (Q632516): Difference between revisions

Let \(\Lambda\) be von Mangoldt's function, \(\mu\) Möbius' function, \(\lambda\) Liouville's function. The authors prove: The Riemann hypothesis is equivalent to each one of the following statements: For every \(f\in C_0^\infty(0,\infty)\), for every \(0<\varepsilon<\frac 12\), \[ \sum_{n=1}^\infty \Lambda(n)f(nx)=\frac{\hat{f}(1)}{x}+o\left(\frac{1}{x^{\frac 12+\varepsilon}}\right), x\to 0, \] where \(\hat{f}\) is the Mellin transform of \(f\). For every \(f\in C_0^\infty(0,\infty)\), for every \(0<\varepsilon<\frac 12\), \[ \sum_{n=1}^\infty \mu(n)f(nx)=o\left(\frac{1}{x^{\frac 12+\varepsilon}}\right), x\to 0. \] For every \(f\in C_0^\infty(0,\infty)\), for every \(0<\varepsilon<\frac 12\), \[ \sum_{n=1}^\infty \lambda(n)f(nx)=o\left(\frac{1}{x^{\frac 12+\varepsilon}}\right), x\to 0. \]