On a new generalization of Hardy-Hilbert's inequality and its applications (Q1290952): Difference between revisions

The main objective of this paper is to prove the following result which generalizes the famous Hardy-Hilbert inequality: \[ \sum^{\infty}_{n, m=0} \frac{a_n b_n}{(m+n+1)^{\lambda}}< B \Big(\frac{p-2+\lambda}{p}, \frac{q-2+\lambda}{q}\Big) \Big[ \sum^{\infty}_{n=0} \frac{a^p_n}{(n+1/2)^{\lambda-1}}\Big]^{\frac 1{p}}\cdot \Big[ \sum^{\infty}_{n=0} \frac{b_n^q}{(n+1/2)^{\lambda-1}}\Big]^{\frac 1{q}}, \] where \(p>1\), \(1/p + 1/q = 1\), \(2-\min \{p,q\}<\lambda \leq 2\), \(a_n, b_n \geq 0\), \(0<\sum^{\infty}_{n=0} (n+1/2)^{1-\lambda}a^p_n\), \(\sum^{\infty}_{n=0} (n+1/2)^{1-\lambda}b^q_n < \infty\), and the symbol \(B\) stands for the \(\beta\) function. Note that the constant factor \(B((p-2+\lambda)/p, (q-2+\lambda)/q))\) in the above mentioned inequality is best possible.