On the hypothesis \(K^*\) of Hardy, Littlewood and Hooley and its relation with discrete fractional operators (Q1000625)

The hypothesis \(K^*\) of Hardy, Littlewood and Hooley states that if \(r_{k}(l)\) denotes the number of ways one can represent the integer \(l\) as the sum of \(k^{\text{th}}\)-powers, \(l= n_1^k + n_2^k + \cdots + n_k^k\), with \(k\) summands then, for every \(\varepsilon>0\), \[ \sum_{l=1}^N r_k(l)^2 = O(N^{1+\varepsilon}) \] as \(N\) tends to \(\infty\) [see \textit{C. Hooley}, Lond. Math. Soc. Lect. Note Ser. 237, 175--185 (1997; Zbl 0930.11071)]. It is shown in [\textit{E. M. Stein} and \textit{S. Wainger}, J. Anal. Math. 80, 335--355 (2000; Zbl 0972.42010)] that the hypothesis \(K^*\) is equivalent to the fact that, for every \(\lambda > \frac{1}{2}\), \[ m_{\lambda, k}(\theta) = \sum_{m=1}^{\infty} \frac{e^{2\pi i m^k \theta}}{m^{\lambda}} \] belongs to \(L^{2k}(\mathbb{T})\). The author gives some new estimates on the function \[ m_{\lambda, \mu}(\theta) = \sum_{m=1}^{\infty} \frac{e^{2\pi i \mu_m \theta}}{m^{\lambda}} \] which may give some new hints on the solution of the hypothesis \(K^*\). Here \(\mu=(\mu_m )_{m}\) is a sequence of positive integers such that, for some positive constants \(A\) and \(B\), \[ Am^{k-1} \leq \mu_m - \mu_{m-1} \leq Bm^{k-1}. \] The author investigates the relation between the hypothesis \(K^*\) and the boundedness of the discrete fractional operator \[ I_{\lambda,k} f(n) = \sum_{m=1}^{\infty} \frac{f(n-m^k)}{m^\lambda} \] with \(k \in \mathbb{N} \setminus \{1\}\), in order to obtain that, for every \(\varepsilon >0\) and every \(2 \leq r < \frac{1}{1-\lambda}\), \[ \int_0^1 |m_{\lambda,k}(\theta)|^r |\theta - a|^{(\frac{1-\lambda}{k} +\varepsilon)r -1} d\theta <\infty \] uniformly in \(a\).