Yahoo Answers is shutting down on May 4th, 2021 (Eastern Time) and beginning April 20th, 2021 (Eastern Time) the Yahoo Answers website will be in read-only mode. There will be no changes to other Yahoo properties or services, or your Yahoo account. You can find more information about the Yahoo Answers shutdown and how to download your data on this help page.
Trending News
For what number a > 1 is there only one solution to a^x = x^a?
I already know the solution. So, I think I can be a fairly good judge of what the best answer is. Show all work in order to get full consideration for the best answer.
The answer is not 2 and not 4. I will tell you why neither of these answers work.
If a = 2, then we would have 2^x = x^2.
In that case, x = 2 or x = 4. After all, 2^2 = 2^2 = 4 and 2^4 = 4^2 = 16.
If a = 4, then would have 4^x = x^4.
Again, the two solutions would be x = 2 and x = 4. Similar to before, 4^2 = 2^4 and 4^4 = 4^4.
In both cases, there were 2 solutions to x. However, I am requiring that there is only 1 solution to x. So, neither a = 2 or a = 4 work.
Hint: In order to solve this problem exactly, you must think beyond just algebra.
6 Answers
- Anonymous1 decade agoFavorite Answer
e=2.7182818...
To see this, consider the function f(t)=t^(1/t). Notice that a^x=x^a if and only if a^(1/a)=x^(1/x), i.e., when f(a)=f(x). Now for positive values of t, f has a maximum value at precisely t=e. You can see this by differentiating f and setting equal to zero. Therefore f(a)=f(x) has two different solutions for x>1, except for x=e.
- PascalLv 71 decade ago
a=e. Consider the function f(x) = x^(1/x). Thus we have:
f(x) = e^(ln x/x)
f'(x) = e^(ln x/x) * (1 - ln x)/x²
We wish to find the maximum of f on the positive real axis. This will occur at a point where f'(x) = 0. Thus we have:
0 = e^(ln x/x) * (1 - ln x)/x²
0 = (1 - ln x)/x²
0 = 1 - ln x
ln x = 1
x = e
For x>e, we have that 1 - ln x is negative, so f'(x) is negative, so f is decreasing. For 0<x<e, 1-ln x is negative, so f'(x) is negative, so f is increasing. This means that x=e is a global maximum of f(x) on the positive real numbers, which means that for all x>0:
f(e) ⥠f(x)
e^(1/e) ⥠x^(1/x)
e^x ⥠x^e
With equality if and only if x = e. So there will be only one solution to the equation e^x = x^e. To show that there is only one such number, let a be any other number greater than 1. Since f(1) = 1 and [xââ]lim f(x) = [xââ]lim e^(ln x/x) = [xââ]lim e^(1/x) = e^0 = 1, it follows from the intermediate value theorem that for any number c strictly between 1 and e^(1/e), there will exist aâ(1, e) and bâ(e, â) such that f(a) = f(b) = c. Further, for any a>1 such that aâ e, f(a) is strictly between 1/e. Which means that if 1<a<e, there exists b such that e<b and f(a) = f(b), and conversely if e<a, there exists b such that 1<b<e and f(a) = f(b).
In either case, if a>1 and aâ e, there exists bâ a such that f(a) = f(b), which means:
a^(1/a) = b^(1/b)
a^b = b^a
However, since a^a = a^a, a and b are thus two distinct solutions to the equation a^x = x^a. Thus, there is one and only one number a>1 for which a^x = x^a has a unique solution in the positive reals, and that is a = e. Q.E.D.
- 1 decade ago
a = e will have only one solution for x iff x is restricted to the Reals, which was not explicitly stated as a requirement. For example if a = e, then one solution is
x â 8.6077123504196488113 - 22.110244684283348950i
another is:
x â -.63000923891390012461 + .48181368120072026243i
There will be many solutions involving the Lambert W function over the Complex numbers. Thinking beyond just algebra....
http://mathworld.wolfram.com/LambertW-Function.htm...
If we are restricted to the Reals, then probably the simplest way to solve this is to consider the function f(x) = (x^a) - (a^x) and look for it's zeros.
Four things are easily proved about this function for a > 1:
f(1) = (1-a) is negative.
Lim(xââ) f(x) = -â is also negative.
f(x = a) is ALWAYS a root of f(x).
d(f(a)) / dx = (aª) * [1 - ln(a)] may be positive, negative or zero.
From these four observations we can reason as follows:
For 1 ⤠x < â:
If f'(a) is positive or negative there must be at least two roots. Thus the only way we can possibly have only one root is if f'(a) is zero. Then:
(aª) * (1 - ln(a)) = 0
ln(a) = 1
a = e
- Anonymous1 decade ago
2
- How do you think about the answers? You can sign in to vote the answer.
- Dr DLv 71 decade ago
To add to what the others have said, here is another way of looking at it.
a^x = x^a implies that
a^(1/a) = x^(1/x) = y
Choosing a value of a that has only one x solution is the same as choosing a value of y that has only one x solution.
Draw the graph of x^(1/x). It has a maximum at x=e (as Pascal has shown).
So draw a horizontal line from y, and you'll see that this line intersects the curve at either
1) 2 places
2) 1 place
3) 0 places
It intersects at only one place if it intersects at the maximum point i.e. when x = a = e.
- 1 decade ago
either 2 or 4:
2^4=4^2
To verify it, the easiest method would be to graph 2^x and x^2, but you probably want a more substantial explanation, so...
in order for two numbers to have and common powers, one must be a power of the other. In order for a^x to equal x^a, when the numbers are written in powers of the smaller number, they equation will be in the form
a^(a^b)=(a^b)^a
which is equivalent to:
a^(a^b)=a^ab
in order for this to be true, a^b must be equal to ab. In order for this to be true, a added to itself b times must be equal to a multiplied by itself b times. For any number greater than 2, the number multiplied by itself b times will be greater than the number added to itself b times. For 2, 2 added twice is the same as 2 multiplied twice and both are equal to 4, so 2 is the solution. However, since the equality can be reversed, 4 is also a solution, so the answers are 2 and 4.