An improved upper bound for the argument of the Riemann zeta-function on the critical line. II (Q2436775)

In the theory of the Riemann zeta-function \(\zeta(s)\), the function \(S(T)\) is defined traditionally by \(\pi S(T)=\arg\zeta(\frac{1}{2}+it)\), where the argument is determined via continuous variation along the line segments connecting \(2\), \(2+iT\) and \(\frac{1}{2}+iT\), with taking the argument of \(\zeta(s)\) at \(s=2\) to be zero. Although conditional approximations of \(S(T)\), assuming the Riemann hypothesis, have the form \[ |S(T)|\leqslant (C+o(1))\frac{\log T}{\log\log T}, \] where \(C\) is an effective constant [\textit{J. E. Littlewood}, Proc. Camb. Philos. Soc. 22, 295--318 (1924; JFM 50.0230.02)], unconditional approximations of \(S(T)\) usually take the form \[ |S(T)|\leqslant a\log T+b\log\log T+c,\quad\text{for \(T\geqslant T_0\)}, \] where \(a\), \(b\), \(c\), and \(T_0\) are computable constants. For over a half of century, the result of \textit{B. Rosser} [Am. J. Math. 63, 211--232 (1941; Zbl 0024.25004)] with \(a=0.137\), \(b=0.443\), \(c=1.588\), and \(T_0=1467\) was the best known in this issue. Recently, the author of present paper [Math. Comput. 81, No. 278, 1053--1061 (2012; Zbl 1267.11093)] improved these constants by obtaining \(a=0.17\), \(b=0\), \(c=1.998\), and \(T_0=\mathrm{e}\). In the paper under review, the author gives a further improvement by proving that \[ a =0.112,\quad b=0.278,\quad c=2.510,\quad\text{and}\quad T_0=\mathrm{e}, \] are admissible. As an immediate corollary, the author obtains an explicit bound for the zero counting function \(N(T)\) which counts the number of zeros \(\beta+i\gamma\) of \(\zeta(s)\) with \(0\leqslant \beta \leqslant 1\), and \(0<\gamma\leqslant T\). The author shows that \[ \left|N(T)-\frac{T}{2\pi}\log\frac{T}{2\pi\mathrm{e}}-\frac{7}{8}\right|\leqslant |S(T)|+\frac{0.2}{T},\quad\text{for \(T\geqslant\mathrm{e}\)}. \]