Unifying inequalities of Hardy, Copson, and others (Q2350254): Difference between revisions

The classical inequalities of Hardy and Copson have been much studied; see for instance [\textit{L.-Y. Chan}, Can. Math. Bull. 22, 165--169 (1979; Zbl 0409.26007); \textit{E. T. Copson}, Proc. R. Soc. Edinb., Sect. A, Math. 75, 157--164 (1976; Zbl 0331.26015); \textit{B. G. Pachpatte}, Indian J. Pure Appl. Math. 23, No. 11, 773--776 (1992; Zbl 0766.26015); \textit{R. N. Mohapatra} and \textit{D. C. Russell}, Aequationes Math. 28, 199--207 (1985; Zbl 0573.26010)]. In this interesting paper, the authors prove three general theorems that give as special cases the classical inequalities mentioned and many of their extensions. An example of the general theorems they prove is Theorem A: If \(0\leq a< x<b\leq \infty\), if \(h\in {\mathcal L}(x,b)\), \(a\leq x\leq b\), \(h\geq 0\), and if \(u>0\), \(U(x) = \int_a^xu< \infty\), \(a\leq x \leq b\), then if \(1<p<\infty\) \[ \int_a^b u(x)\Bigl(\int_x^bh\Bigr)\text{d}x\leq p^p\int_a^bu\Bigl({hU \over u}\Bigr)^p, \] where either side may be infinite. If \(0<p<1\) the inequality is reversed and if \(p=1 \) there is equality. Theorem B is a similar result with the interval \((x,b)\) replaced by \((a,x)\). These results are related to a lemma used by \textit{G. S. Davies} and \textit{G. M. Petersen} in [Q. J. Math., Oxf. II. Ser. 15, 35--40 (1964; Zbl 0138.03406)]. Finally, there is Theorem C that is related to the Levinson extension of Hardy's inequality. The authors point out that their very wide umbrellas do not seem to cover the extensions of \textit{E. R. Love} [J. Lond. Math. Soc., II. Ser. 30, 431--440 (1984; Zbl 0526.26010); Proc. R. Soc. Edinb., Sect. A, Math. 100, 237--262 (1985; Zbl 0584.26011)].