Find the number of ordered pairs of real numbers ( x , y ) such that
{ x = y + 1 0 0 0 − y + 1 0 0 y = x + 1 0 0 0 − x + 1 0 0 .
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
Sir i solved in this way
y + 1 0 0 0 = a , y + 1 0 0 = b
a − b = 9 0 0
( y + 1 0 0 0 + y + 1 0 0 ) ( y + 1 0 0 0 − y + 1 0 0 ) = 9 0 0
y + 1 0 0 0 − y + 1 0 0 = x 9 0 0
adding ,
2 y + 1 0 0 0 = x 9 0 0 + x
and similarly for the other equation , we get
2 x + 1 0 0 0 = y 9 0 0 + y
thus both the curves can only intersect at point where x=y
Nice.
Great solution.Sir , you are Brilliant
Let f ( t ) = t + 1 0 0 0 − t + 1 0 0 . We can write
f ( t ) = t + 1 0 0 0 − t + 1 0 0 = t + 1 0 0 0 + t + 1 0 0 ( t + 1 0 0 0 − t + 1 0 0 ) ( t + 1 0 0 0 + t + 1 0 0 ) = t + 1 0 0 0 + t + 1 0 0 ( t + 1 0 0 0 ) − ( t + 1 0 0 ) = t + 1 0 0 0 + t + 1 0 0 9 0 0 .
In this form, we can see that the function f is decreasing.
We want to find x and y such that x = f ( y ) and y = f ( x ) .
If x < y , then f ( x ) > f ( y ) . But x = f ( y ) and y = f ( x ) , so f ( y ) > f ( x ) , contradiction. Similarly, x > y also leads to a contradiction. Therefore, if x = f ( y ) and y = f ( x ) , then x = y . In other words, f ( x ) = x .
Consider the continuous function g ( x ) = f ( x ) − x . We see that g ( 0 ) = f ( 0 ) = 1 0 0 0 − 1 0 0 > 0 , and
g ( 2 4 ) = f ( 2 4 ) − 2 4 = 1 0 2 4 − 1 2 4 − 2 4 = 8 − 1 2 4 < 0 ,
so by the Intermediate Value Theorem, there exists an x 0 ∈ [ 0 , 2 4 ] such that g ( x 0 ) = 0 , or f ( x 0 ) = x 0 . Furthermore, the function g ( x ) = f ( x ) − x is decreasing, so x = x 0 is the unique solution of f ( x ) = x .
Hence, the system x = f ( y ) and y = f ( x ) has exactly one solution, namely x = y = x 0 .
I anticipated that someone would create a solution along these lines. It is not true that if x = f ( y ) , y = f ( x ) for a decreasing function f ( x ) , then the only solutions must be of the form ( x , y ) = ( α , α ) . Easy counterexamples to this statement are the functions f ( x ) = − x [which has solutions of the form ( α , − α ) ] and f ( x ) = x 1 [which has solutions of the form ( α , α 1 ) ].
I am sorry, but I don't follow. It is clear that f(x) is decreasing but where is the contradiction? If x < y then f ( x ) > f ( y ) and from the equations we get the same not the opposite.
Log in to reply
Indeed. This is a false solution, there is no contradiction.
The 'correct' version is that if f ( x ) is an increasing function, then x = f ( y ) and y = f ( x ) only has solutions x = y . This can be proved in the manner stated above, namely that if x > y then by assumption of increasing function we get that f ( x ) > f ( y ) and hence y > x which is a contradiction.
Log in to reply
I can't believe I made such an obvious mistake. Sorry to everyone who was misled.
Here's my solution.
Let f ( t ) = t + 1 0 0 0 − t + 1 0 0 for t ≥ − 1 0 0 as before. I'll list out the steps to arrive at a contradiction for y = f ( x ) , x = f ( y ) , x = y .
On the other hand, let g be the inverse of f on (0, 30); note that g ′ ( t ) = f ′ ( g ( t ) ) 1 . We then have y = g ( x ) , x = g ( y ) , x = y and:
Now we'll get a contradiction: since f ( t ) = g ( t ) has two distinct solutions 2 0 ≤ x , y ≤ 2 2 , Rolle's theorem tells us there must be a point 2 0 ≤ u ≤ 2 2 such that f ′ ( u ) = g ′ ( u ) . But this is clearly impossible.
taking a cue from Jon's answer I did this:
(sqrt(x+1000) + sqrt(x+100)) / x = (sqrt(y+1000) + sqrt(y+100)) / y = 900 / xy.
This shows that signs of x and y are same, either both positive or both negative.
Now the function f(t) = (sqrt(t+1000) + sqrt(t+100)) / t is strictly decreasing for t>0, and strictly decreasing for t<0. At t = 0 it is undefined, but we can check that case separately. We want f(x) = f(y), which is not possible for different x,y in a monotonous function. So x=y is the only possibility.
Log in to reply
just rewriting in a bit different way :)
Setting f ( t ) = t + 1 0 0 0 − t + 1 0 0 ,We have x f ( x ) = y f ( y ) .But the function g ( t ) = t f ( t ) = t ( t + 1 0 0 0 − t + 1 0 0 ) is increasing as t > 0 .Since both x and y are positive,then they have to be equal
The given equations are symmetric about the line y = x as you might have noticed. Anyway, good algebraic interpretation & nice use of IVT to ascertain the presence of the root.
Log in to reply
All that implies is that f ( f ( x ) ) = x , and it doesn't necessarily imply that f ( x ) = x .
Of course, f ( x ) = x ⇒ f ( f ( x ) ) = x . It is the reverse implication that we need, and it is not necessarily true.
Log in to reply
For some (unknown) reason, I can not directly post the solution there. So I am outlining it here. First of all it is clear that we can concentrate on positive real values of x and y only and the relevant equation to investigate is f ( f ( x ) ) = x where f ( x ) = x + 1 0 0 0 − x + 1 0 0 . Next the important step is to show that the mapping f : R + → R + is a contraction . This is easy as the absolute value of derivative of f which is = 2 1 ( x + 1 0 2 1 − x + 1 0 3 1 ) is upper-bounded by 2 1 1 0 1 < 1 . This implies that the mapping f ( f ( x ) : R + → R + is also a contraction. Now use Banach Fixed Point Theorem for the existence and uniqueness of the fixed point of f ( f ( x ) ) = x .
Log in to reply
@Abhishek Sinha – Beautiful - it didn't occur to me to use Banach fixed point theorem.
Log in to reply
@C Lim – These Fixed Point Theorems are one of the most powerful (as well as neat) tools to show existence (and sometimes uniqueness, if the relevant hypotheses are satisfied) of fixed points of certain kinds of self-mappings even in a general metric space (subject to, of course, necessary regularity conditions).
@Abhishek Sinha – Very nice.
When I originally posted the inequality:
d x d [ f ( f ( x ) ) − x ] = f ′ ( f ( x ) ) f ′ ( x ) − 1 < 4 0 0 1 − 1
I could see that that would do the trick (in particular, f ( f ( x ) ) − x cannot equal zero more than once, if it is strictly decreasing) but I didn't have much intuitive sense of the underlying structure. Now I do though, after reading your solution, Abhishek -- as you said, if f ( x ) is a contraction, then so is f ( f ( x ) ) .
On a side note, I see that you, I, and C L all used calculus ( ∣ f ′ ( x ) ∣ < 2 0 1 < 1 or a variation of that). It may go without saying, but I'll say it anyhow, that we could avoid calculus if we wanted to:
f ( y ) − f ( x )
= y + 1 0 0 0 − x + 1 0 0 0 + x + 1 0 0 − y + 1 0 0
= y + 1 0 0 0 + x + 1 0 0 0 y − x − y + 1 0 0 + x + 1 0 0 y − x
= ( y − x ) ⋅ G ( x , y )
with
But − 2 0 1 ≤ G ( x , y ) ≤ 0 for any x , y ≥ 0 , which gives us what we want.
Log in to reply
@Peter Byers – That's true. In fact, I would rather prefer to avoid the use of calculus in this example (albeit, at the expense of a few more lines). Because the solution has got nothing to do with differentiablity of R per se. It is a purely topological property.
@Peter Byers – As far as I can tell, these solutions doesn't generalize to versions where we replace 1000 and 100 with other numbers, because they rely on a 'small rate of change' argument.
For example, in Peter's new solution, to show that G ( x , y ) ≥ − 1 , we can't replace 1 0 0 with a value smaller than 4 1 (and replace 1000 with a large enough number).
Can you find a solution which works when 1 0 0 0 , 1 0 0 are replaced with any 2 positive numbers?
Jon, ...really fabulous and good explanatory solution....
I solved this equation graphically
It was evident that due to symmetry, the solution would lie on x = y and it is evident that x > = − 1 0 0 and y > = − 1 0 0
So I wrote a MATLAB code as follows :
x 1 = [ − 1 0 0 : 2 0 0 ] ;
y 1 = s q r t ( x 1 + 1 0 0 0 ) − s q r t ( x 1 + 1 0 0 ) ;
y 2 = x 1 = [ − 1 0 0 : 2 0 0 ] ;
x 2 = s q r t ( y 2 + 1 0 0 0 ) − s q r t ( y 2 + 1 0 0 ) ;
%Time to plot
p l o t ( x 1 , y 1 ) ;
h o l d a l l ;
p l o t ( x 2 , y 2 ) ;
voila! As expected, the point lies on x = y and there exist only 1 such point
P.S. I myself dont support graphical solutions much, so I apologize if I offend any one of you. After Spending much time on the question, out of curiosity, I chose the Graphical Method ^_^
Aman G. wrote:
I agree.. Paramjit. A Graphical Solution can hardly qualify as a proper solution.
In my case, I didn't rely on a graphical method, but I did rely on calculus. Which is to say, my solution could be considered a "proper mathematical statement" but not the one intended by the staff. I'll describe it anyhow: we can show that for any t > 0
− 1 / 2 0 < f ′ ( t ) < 0
from which it follows that
d x d [ f ( f ( x ) − x ] = f ′ ( f ( x ) ) f ′ ( x ) − 1 < 4 0 0 1 − 1 < 0
and so on.
Log in to reply
Great job! You identified a (sufficient) condition, namely that if − 1 < f ′ ( t ) < 1 , then f ( f ( x ) ) = x ⇒ f ( x ) = x .
This shows that the important graphical condition is that the graph of f ( t ) doesn't change too abruptly. Without this explanation, Aman's solution would be incomplete.
It is easy to see how the examples that I gave, i.e. f ( x ) = − x , f ( x ) = x 1 , do not satisfy your condition.
Log in to reply
Thanks Calvin!
I like your way of putting it ... among other things, that gives us the option of removing the chain rule (and compositions of functions) from the solution: If b = f ( a ) = a = f ( b ) then the average rate of change on the interval [ a , b ] is b − a f ( b ) − f ( a ) = b − a a − b = − 1
which cannot happen if f ′ > − 1 on the same interval.
Try to learn from Jon's solution above. I was once myself in your place, but these great people have corrected me! :)
I agree.. Paramjit. A Graphical Solution can hardly qualify as a proper solution. Also, it can never match the beauty of a analytical solution. Actually you know.. I was able to come up with the idea of rationalising , but got stuck after that :D (the idea of decreasing function did not strike me)
It is not true that "due to symmetry" the solution must lie on the line y = x .
For example, solve the system of equations
{ y = x 1 x = y 1
Are there solutions other than ( 1 , 1 ) and ( − 1 , − 1 ) ?
Of course, your graphical solution would indeed show that there is only 1 solution, and that my above system has infinitely many solutions. Can you take this idea and make it into a proper mathematical statement?
Log in to reply
I can't post an official solution because I wanted to see how others did it, so I will do it as a comment to thing that is as close as what I did. I am not sure about this. I will try to explain how I found out it only had 1 solution.
Let's consider these functions :
x = y y = x
The solutions lie on the line x=y because it the same function rotated 90 degrees, and because it's centered at the origin, the symmetry line is y=x. We have ( (0,0) and (1,1) ) without considering the imaginary part.
Let's try to get another definition closer to to the problem.
x = y + a − y + b y = x + a − x + b
This is the exact same function except x changed for y. if a > b , the result of the function is always greater than 0 (it's a pair root). The domain of the function is, if we use the restriction, is always : D o m x : [ − b , + ∞ [
for y(x) function and for the x(y) we simply have the same domain D o m y : [ − b , + ∞ [
This should at least admit 1 solution, which is the point where they cross on y=x. The problem is that the function y is not the reciprocal of x, but they behave like a reciprocal function. If we find the point where the function starts, we have :
y ( − b ) = a − b x ( − b ) = a − b
In the context it gives us y ( − 1 0 0 ) = 3 0 x ( − 1 0 0 ) = 3 0
The symmetry between these 2 points is y=x,Let's try another random point to see if they are symmetrical to y=x.
y ( 0 ) = a − b = 1 0 0 0 − 1 0 0 x ( 0 ) = a − b = 1 0 0 0 − 1 0 0
The last thing to do is find all the points where the function cross y=x, if a > b , there is only 1 solution to the equation and it's approximately
( 2 0 . 9 5 4 4 4 ; 2 0 . 9 5 4 4 4 )
If there is anything wrong in what I wrote I would really like to know why, I find it's a very interesting problem.
Log in to reply
I do not see any arguments presented for why the only solution must occur at x = y .
I could replace it with x = − y 1 , and have all your statements still be true, but the conclusion is clearly false.
Log in to reply
@Calvin Lin – If all the points of the function are symmetrical to y=x, the solutions, if they exist, must be also symmetrical, and to have a solution, (x,y) must be equal to (y,x). As shown with the square of x or y. The function always increases. A basic square root function admit 2 solutions, the modified function admit 1 solution and because all the points are symmetrical,t must be y=x
Log in to reply
@Samuel Hatin – Please read the other comments.
Your first statement is incorrect. If all the points of the function are symmetrical to y = x , this means that f ( f ( x ) ) = x for all values of x , and hence we have many solutions of the form ( α , f ( α ) ) , and they do not need to be of the form ( α , α ) which is what you claim.
An example I gave was f ( x ) = − x , in which we do have f ( f ( x ) ) = f ( − x ) = x and hence points of the form ( α , − α ) satisfy the system of equations that f ( α ) = − α , f ( − α ) = α ) .
It can be seen that no matter how, it is impossible for x and y to take on distinct values. Hence, x=y and we have x= \sqrt{x+1000} - \sqrt{x+100}. Solving this, we get a quartic equation with four roots, x= -37.4, -22.5, 21.0 and 44.3. However, three of them are invalid. The proof is as follows: One must observe that since the square root of any number is positive, \sqrt{x+1000} > \sqrt{x+100} for all x. Hence, x must be positive and we reject the two negative roots. Manipulating the equation gives us x+ \sqrt{x+100} = \sqrt{x+1000} Squaring it gives us x^2 + 2x\sqrt{x+100} =900 then 2x\sqrt{x+100} = 900 -x^2. Since 2x\sqrt{x+100} >0, 900- x^2 >0 and we have 0<x<30. Hence we reject x= 44.3 leaving us with only one root and hence one ordered pair.
x + 1 0 0 is a curve (which looks like half of a parabola, like x ) lying in the first quadrant starting from the point { 0 , 1 0 0 } . Similarly x + 1 0 0 0 has the same structure but starts from { 0 , 1 0 0 0 } . Since both functions are monotonic increasing, their difference is also monotonic increasing, and hence y = x + 1 0 0 0 − x + 1 0 0 is a curve of similar shape but passes between the two individual curves and starts from { 0 , 1 0 0 0 − 1 0 0 } . Now for the curve x = y + 1 0 0 0 − y + 1 0 0 , it is simply a rotated version (also lying in the first quadrant and starting from the point { 1 0 0 0 − 1 0 0 , 0 } and heading towards y = infinity. Clearly, both the curves meet at a single point. And hence the solution is 1
We can easily see that (just by looking at one of the two roots in each equation) we have two identical parabola (half of them) with only one point of intersection on the line y=x Then we only take the same quantity from each of them, so the intersection will remain one
(Strictly speaking,) It is not true that we have a parabola. In fact, you can show that y = 4 x 2 x 4 − 2 2 0 0 x 2 + 8 1 0 0 0 0 , which has some similarities to a parabola when ∣ x ∣ is very large, but behaves very differently when ∣ x ∣ is close to 0.
Having only 1 point of intersection on the line y = x is not relevant to this question. You are missing out a crucial detail.
This is not rigours solution, but it's how I guess the answer
y = x + 1 0 0 0 − x + 1 0 0 , f ′ ( x ) < 0 where − 1 0 0 ≤ x ≤ ∞ , Stricly Decreasing.
x = y + 1 0 0 0 − y + 1 0 0 , f ′ ( y ) < 0 where − 1 0 0 ≤ y ≤ ∞ , this graphic same with rotation fitst graphic irotation 1 8 0 ∘ counterclockwise.
Problem Loading...
Note Loading...
Set Loading...
Most of the solutions presented here are wrong, because they didn't provide an adequate explanation of why the only solutions must occur at x = y . This was either assumed, or incorrectly shown.
Comments posted by Peter B., Abhishek S. and C L. provides solutions based on calculus techniques like mean value theorem, contraction mapping, Rolle's theorem.
I would encourage you to read through the long thread of discussions, and learn from each other.
Peter presents the following solution which doesn't use Calculus.
Let f ( x ) = x + 1 0 0 0 − x + 1 0 0 . We are given that f ( x ) = y and f ( y ) = x . Observe that
f ( y ) − f ( x ) = y + 1 0 0 0 − x + 1 0 0 0 + x + 1 0 0 − y + 1 0 0 = y + 1 0 0 0 + x + 1 0 0 0 y − x − y + 1 0 0 + x + 1 0 0 y − x = ( y − x ) ⋅ G ( x , y )
where G ( x , y ) = y + 1 0 0 0 + x + 1 0 0 0 1 − y + 1 0 0 + x + 1 0 0 1
But since − 2 0 1 ≤ G ( x , y ) ≤ 0 for any x , y ≥ 0 , in order for y − x = f ( y ) − f ( x ) , we must have y = x . We then proceed to solve the equation, and show that only 1 solution exists.
The following is my solution.
Multiplying by the algebraic conjugate, the system is equivalent to
{ x ( y + 1 0 0 0 + y + 1 0 0 ) = 9 0 0 y ( x + 1 0 0 0 + x + 1 0 0 ) = 9 0 0 .
As such, x and y must be positive. If x > y , then we have
9 0 0 = x 2 y + 1 0 0 0 x 2 + x 2 y + 1 0 0 x 2 > x y 2 + 1 0 0 0 y 2 + x y 2 + 1 0 0 y 2 = 9 0 0 ,
which is a contradiction. A similar inequality holds for x < y , which means we must have x = y .
Thus, x ( x + 1 0 0 0 + x + 1 0 0 ) = 9 0 0 . Observe that the LHS is an increasing function on positive values of x , and is equal to 0 when x = 0 and increases without bound. Hence, exactly 1 solution exists.
Note: This easily generalizes to other (positive) values. When N is negative, we can't just substitute x = 0 in. Instead, the smallest starting value is x = − N , which gives us a value of ∣ N ∣ M + ∣ N ∣ .