3
$\begingroup$

Let $a<b$ be positive real numbers , then is it true that there exist $x,y \in (a,b)$ such that

$ \bigg(\dfrac { a+b}2\bigg)^{x+y}=a^xb^y$ ?

$\endgroup$
0
1
$\begingroup$

Fix $a\lt b$ positive. The function $F$ defined by $F(u)=a^{1-u}b^{u}$ for every $u$ in $[0,1]$ increases from $F(0)=a$ to $F(1)=b$. By continuity, and since $c=\frac12(a+b)$ is in $(a,b)$, there exists some $u(c)$ in $(0,1)$ such that $F(u(c))=c$. Every $(x,y)$ such that $u(c)x=(1-u(c))y$ solves $c^{x+y}=a^xb^y$ hence it remains to see whether one such couple $(x,y)$ is such that $x$ and $y$ are in $(a,b)$.

This happens if and only if, when $x=a$ and $u(c)x=(1-u(c))y$ then $y\lt b$, that is, if and only if $u(c)a\lt(1-u(c))b$, or, equivalently, $u(c)\lt b/(a+b)$, or, equivalently, $c\lt F(b/(a+b))$, that is, $c^{a+b}\lt a^ab^b$, that is, finally, $(a+b)\log c-a\log a-b\log b\lt0$.

Fix some $a\gt0$ and, for every $b\geqslant a$, consider $G(b)=(a+b)\log\left(\frac12(a+b)\right)-a\log a-b\log b$, then $G(a)=0$ and $G'(b)=\log\left(\frac12(a+b)\right)-\log b\lt0$ for every $b\gt a$ hence $G(b)\lt0$ for every $b\gt a$, QED.

$\endgroup$
1
$\begingroup$

We can solve the equation by $$(x(t),y(t)) = \left(\ln\left(\frac{2b}{a+b}\right)t,\ln\left(\frac{a+b}{2a}\right)t\right)$$

Now $\frac{a+b}{2}\geq a$ and therefore $\ln\left(\frac{a+b}{2a}\right)\geq 0$ and by $\frac{a+b}{2}\leq b$ we conclude $\ln\left(\frac{2b}{a+b}\right)\geq 0$.

Furthermore $\ln\left(\frac{a+b}{2a}\right) \geq \ln\left(\frac{2b}{a+b}\right)$

Thus for $$t \in \left(\frac{a}{\ln\left(\frac{2b}{a+b}\right)},\frac{b}{\ln\left(\frac{a+b}{2a}\right)}\right)$$ we have $x(t),y(t)\in (a,b)$.

$\endgroup$
0
0
$\begingroup$

Another way: With $t=\frac{x}{x+y}$, we seek solutions for $$\frac{a+b}2=a^tb^{1-t}$$ With the RHS being continuous, for solutions to exist in the right interval, it is sufficient to show that $$a^\frac{b}{a+b}b^\frac{a}{a+b}<\frac{a+b}2<a^\frac{a}{a+b}b^\frac{b}{a+b}$$

The left inequality is a consequence of weighted AM-GM, the right one follows on taking logs and using Jensen's inequality.

$\endgroup$
0
$\begingroup$

There should be an infinite number of solutions.

If we let x = f (t), y = g (t) for two continuous functions on [0, 1] with f (0) = 1, g (0) = 0, and f (1) = 0, g (1) = 1, then the left hand side of the equation is (a + b) / 2 at t = 0 and t = 1, while the right hand side changes from a < (a + b)/2 continuously to b > (a + b)/2 when t goes from 0 to 1. So for some 0 < t < 1 both sides must be equal. And since we picked f, g with very little restrictions, there should be a huge set of solutions.

Oops. Just found the restriction a <= x, y <= b.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.