Difference between revisions of "Callebaut's Inequality"
Spanferkel (talk | contribs) m (→Proof) |
Spanferkel (talk | contribs) m (→Proof) |
||
Line 15: | Line 15: | ||
Raising these three respectively to the <math>(\frac xy-1)</math>th, <math>\frac xy</math>th, <math>\frac xy</math>th power, we get | Raising these three respectively to the <math>(\frac xy-1)</math>th, <math>\frac xy</math>th, <math>\frac xy</math>th power, we get | ||
− | |||
− | |||
− | + | <math>f(1+y)^{\frac xy-1}\cdot f(1-y)^{\frac xy-1} \ge f(1)^{2(\frac xy-1)}</math> | |
− | f(1- | ||
− | \ | + | <math>f(1-x)\cdot f(1)^{\frac xy-1} \ge f(1-y)^{\frac xy}</math> |
− | f(1+x)\cdot f(1)^{\frac xy-1} \ge f(1+y)^{\frac xy} | + | |
+ | <math>f(1+x)\cdot f(1)^{\frac xy-1} \ge f(1+y)^{\frac xy}</math> | ||
Multiplying the last three lines yields <math>f(1+x)f(1-x)\ge f(1+y)f(1-y)</math> as required. | Multiplying the last three lines yields <math>f(1+x)f(1-x)\ge f(1+y)f(1-y)</math> as required. |
Latest revision as of 16:30, 4 December 2010
Callebaut's Inequality states that for
It can be considered as an interpolation or a refinement of Cauchy-Schwarz, which is the case.
Proof
Let . Then by Hölder,
, further (because )
and
.
Raising these three respectively to the th, th, th power, we get
Multiplying the last three lines yields as required.