Currently displaying 1 – 5 of 5

Showing per page

Order by Relevance | Title | Year of publication

The mean square of the divisor function

Chaohua JiaAyyadurai Sankaranarayanan — 2014

Acta Arithmetica

Let d(n) be the divisor function. In 1916, S. Ramanujan stated without proof that n x d ² ( n ) = x P ( l o g x ) + E ( x ) , where P(y) is a cubic polynomial in y and E ( x ) = O ( x 3 / 5 + ε ) , with ε being a sufficiently small positive constant. He also stated that, assuming the Riemann Hypothesis (RH), E ( x ) = O ( x 1 / 2 + ε ) . In 1922, B. M. Wilson proved the above result unconditionally. The direct application of the RH would produce E ( x ) = O ( x 1 / 2 ( l o g x ) l o g l o g x ) . In 2003, K. Ramachandra and A. Sankaranarayanan proved the above result without any assumption. In this paper, we prove E ( x ) = O ( x 1 / 2 ( l o g x ) ) .

On the average behavior of the Fourier coefficients of j th symmetric power L -function over certain sequences of positive integers

Anubhav SharmaAyyadurai Sankaranarayanan — 2023

Czechoslovak Mathematical Journal

We investigate the average behavior of the n th normalized Fourier coefficients of the j th ( j 2 be any fixed integer) symmetric power L -function (i.e., L ( s , sym j f ) ), attached to a primitive holomorphic cusp form f of weight k for the full modular group S L ( 2 , ) over certain sequences of positive integers. Precisely, we prove an asymptotic formula with an error term for the sum S j * : = a 1 2 + a 2 2 + a 3 2 + a 4 2 + a 5 2 + a 6 2 x ( a 1 , a 2 , a 3 , a 4 , a 5 , a 6 ) 6 λ sym j f 2 ( a 1 2 + a 2 2 + a 3 2 + a 4 2 + a 5 2 + a 6 2 ) , where x is sufficiently large, and L ( s , sym j f ) : = n = 1 λ sym j f ( n ) n s . When j = 2 , the error term which we obtain improves the earlier known result.

On the Riesz means of n/ϕ(n) - III

Ayyadurai SankaranarayananSaurabh Kumar Singh — 2015

Acta Arithmetica

Let ϕ(n) denote the Euler totient function. We study the error term of the general kth Riesz mean of the arithmetical function n/ϕ(n) for any positive integer k ≥ 1, namely the error term E k ( x ) where 1 / k ! n x n / ϕ ( n ) ( 1 - n / x ) k = M k ( x ) + E k ( x ) . For instance, the upper bound for |Ek(x)| established here improves the earlier known upper bounds for all integers k satisfying k ( l o g x ) 1 + ϵ .

Page 1

Download Results (CSV)