P = ( x + y ) ( z + t )
Let x , y , z , t be positive real numbers satisfying x 2 + y 2 + z 2 + t 2 = 1 0 and x y z t = 4 .
If the maximum value of the expression is S , find ⌊ 1 0 0 S ⌋ .
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.
To simplify the equations, I started with a linear substitution x = a + b , y = a − b , z = c + d , t = c − d . We are now asked to maximize P 2 = 1 6 a 2 c 2 , subject to the constraints ( a 2 − b 2 ) ( c 2 − d 2 ) = 4 and a 2 + b 2 + c 2 + d 2 = 5 . Eliminating c and d from these equations, we are asked to maximize P 2 = 1 6 a 2 ( 2 5 − a 2 + 2 a 2 − b 2 + a 2 − b 2 2 ) . For a fixed a , there are no maxima in the interior; consider the two terms involving b , which are reciprocals. Thus the maximum must be attained at a boundary point, where b = 0 or d = 0 . By symmetry, we can assume that b = 0 , with P 2 = 8 2 − 8 ( a 2 − 5 / 2 ) 2 and S = 8 2 when a 2 = 5 / 2 . One point where this maximum is attained is x = y = 5 / 2 , z = 2 0 4 1 + 3 , t = 2 0 4 1 − 3 .
Simply beautiful.
Very nice solutions all! This problem was tricky. Thanks for sharing! (Even if I'm having trouble understanding Otto Bretscher's solution.) I searched your name on the internet by the way, and found your textbook on linear algebra. Then, after that, I was watching Benedict Gross lecture on abstract algebra, and he mentioned your book! I do hope to be able to read it at some point in the future. I should be able to afford it eventually.
Oh that's nice. I thought of P 2 , but was unable to push through.
For clarity, the equality condition requires x 2 + y 2 = z 2 + t 2 and x z + y t = x t + y z . Thus, we have x 2 + y = z 2 + t 2 = 5 and ( x − y ) ( z − t ) = 0 .
[This is not a complete solution, but it sheds some insight into why the answer isn't just 900.]
For those who think that the answer is 1000, you forgot to account for x y z t = 4 , which isn't satisfied when x = y = z = t = 2 5 .
Most people think that the maximum occurs when { x , y } = { z , t } = { 1 , 2 } (in part because of how nice it looks). While this set of values do satisfy the conditions, they do not maximize the product P .
Note: I used Lagrange multipliers to solve this problem, which has a boundary condition (which is partly what makes this question interesting). You should see Khang's solution for a classical approach.
Instead, let's set a = x 2 + y 2 and b = x y . We obtain 1 0 − a = z 2 + t 2 and b 4 = z t . We then want to find the maximum of
P 2 = ( x 2 + y 2 + 2 x y ) ( z 2 + t 2 + 2 z t ) = ( a + 2 b ) ( 1 0 − a + b 8 ) .
We are subject to
x
2
+
y
2
+
z
2
+
t
2
≥
x
2
+
y
2
≥
2
x
y
≥
0
, or that
1
0
≥
a
≥
2
b
≥
0
and
x
2
+
y
2
+
z
2
+
t
2
≥
z
2
+
t
2
≥
2
z
t
≥
0
, or that
1
0
−
a
≥
b
8
.
Claim: This is maximized at
{
(
a
,
b
)
}
=
{
(
5
,
2
.
5
)
}
or
{
(
5
,
1
.
6
)
}
.
I used Lagrange multipliers to show this. We are at a boundary condition.
We have equality when { { x , y , } , { z , t } } = { { 2 . 5 , 2 . 5 } , { ≈ 0 . 7 6 1 , ≈ 2 . 1 } } .
The maximum value of P = 8 2 .
@Khang Nguyen Thanh I am interested in seeing your solution. Care to share?
S i r , p l e a s e c a n y o u t e l l w h e r e w a s I w r o n g , I d i d i t a s : 2 ( x + y ) + ( z + t ) ≥ ( x + y ) ( z + t ) S q u a r i n g b o t h s i d e s : ( x + y ) 2 + ( z + t ) 2 + 2 ( x + y ) ( z + t ) ≥ 4 P x 2 + y 2 + z 2 + t 2 + 2 x y + 2 z t + 2 P ≥ 4 P 1 0 + 2 ( x y + z t ) ≥ 2 P 5 + ( x y + z t ) ≥ P N o w P w i l l b e m a x . w h e n x y + z t i s m a x . A l s o b y A . M . ≥ G . M . x 2 + y 2 ≥ 2 x y a n d z 2 + t 2 ≥ 2 z t o r 1 0 ≥ 2 ( x y + z t ) ∴ x y + z t ≤ 5 ∴ P m a x = 5 + 5 = 1 0 . S = 1 0 ; 1 0 0 S = 1 0 0 0 . P l e a s e t e l l w h e r e w a s I w r o n g .
Log in to reply
All that you have found is an upper bound. You need to show that it is the lowest upper bound , in order to conclude that it is the maximum. IE all that you have is P max ≤ 1 0 , and you have not demonstrated that equality can exist (which it cannot).
Note: In addition, you did not use the condition that x y z t = 4 . In your case, equality holds when x + y = z + t , x = y , z = t , which forces x = y = z = t = 5 2 , and hence x y z y = 5 = 4 .
This was exactly what I thought. My solution was the same as Samarth's except, at the end, when I got to P ≤ 5 + x y + z t with x + y = z + t (to produce equality with the GM-AM), I just assumed there was probably another inequality I could've used, and I would've ended up setting x y = z t .This was obviously a bad guess.
Sir what if I get 1000 with another method? Can u tell me where I was wrong? P=(X+y)(z+t) =Xz+xt+yz+yt ≤(xx+zz)/2+(xx+tt)/2+(yy+zz)/2+(yy+tt)/2(applying am gm) ≤2(xx+yy+zz+tt)/2 ≤10(Pmax) Hence S=10,100S=1000
Log in to reply
As always, when asked to show that something is a maximum, you have to
1. Show that it is an upper bound
2. Show that it is the least upper bound, aka that it can be achieved.
You have shown that 10 is an upper bound for S. What are the conditions for equality to hold throughout? For what values does equality hold?
Note: In particular, you didn't use the condition that x y z t = 4 , so that should be a hint that your solution is incomplete.
First, I let x y = a (noting a > 0 ) so that z t = a 4 , by the equation x y z t = 4 . Then I let x + y = r . In addition, I applied the simplifying assumption x > y , which can be justified by symmetry. Then I manipulated x 2 + y 2 + z 2 + t 2 = 1 0 in the following way: x 2 + 2 x y + y 2 + z 2 + 2 z t + t 2 = 1 0 + 2 a + a 8 ⇒ ( z + t ) 2 = 1 0 + 2 a + a 8 − r 2 ⇒ P 2 = r 2 ( 1 0 + 2 a + a 8 ) − r 4 Next, I determined the feasible region in the a r -plane. The equation a = x y alone places only the restriction a > 0 on the value of a . Next, substitute t = a z 4 in x 2 + y 2 + z 2 + t 2 = 1 0 to obtain: 1 0 = x 2 + y 2 + z 2 + a 2 z 2 1 6 ≥ x 2 + y 2 + a 8 = r 2 − 2 a + a 8 Note that the inequality, r 2 − 2 a + a 8 ≤ 1 0 , produced is a sharp inequality (consider z = a 2 ). Next, note that, because x = 2 r + r 2 − 4 a and y = 2 r − r 2 − 4 a , we must have r ≥ 2 a because the quantity inside the square root operator must be nonnegative. Combining these two findings gives the region of interest in the a r -plane: 2 a ≤ r ≤ 1 0 + 2 a − a 8 Setting the first partial derivates of P 2 = r 2 ( 1 0 + 2 a + a 8 ) − r 4 w.r.t. r and a equal to zero, I determine there are no local extrema in the region of interest. This means that the maximum must lie on the boundary. Therefore, I test the two boundary curves, r = 1 0 + 2 a − a 8 and r = 2 a , both with 1 ≤ a ≤ 4 . On the boundary curve r = 1 0 + 2 a − a 8 , P 2 obtains its maximum of 7 6 + 9 4 when a = 5 1 2 . Along the other boundary curve r = 2 a , P 2 obtains a maximum of 8 2 when a = 2 . 5 . Therefore, the maximum value of P , subject to the constraints, is 8 2 .
Problem Loading...
Note Loading...
Set Loading...
By using AM-GM inequality, we have:
P = ( x + y ) ( z + t )
= ( x 2 + y 2 + 2 x y ) ( z 2 + t 2 + 2 z t )
= ( x 2 + y 2 ) ( z 2 + t 2 ) + 2 x y ( z 2 + t 2 ) + 2 z t ( x 2 + y 2 ) + 4 x y z t
≤ ( 2 x 2 + y 2 + z 2 + t 2 ) 2 + 2 ( x z + y t ) ( x t + y z ) + 1 6
≤ 2 5 + 2 ( 2 x z + y t + x t + y z ) 2 + 1 6
= 4 1 + 2 1 P 2
Solving this inequality gives us that P ≤ 8 2 .
The equality conditions are:
1. x 2 + y 2 = z 2 + t 2 = 5
2. x y z t = 4
3. ( x + y ) ( t + z ) = 8 2 .
Squaring and expanding condition 3, and substituting in with conditions 1 and 2, we get ( 5 + 2 x y ) ( 5 + x y 8 ) = 8 2 , or that x y = 2 5 , 5 8 .
Thus, solving x 2 + y 2 = 5 , x y = 2 5 ⇒ { x , y } = { 2 5 , 2 5 } and z 2 + t 2 = 5 , z t = 5 8 ⇒ { z , t } = { 2 5 + 5 1 6 + 5 − 5 1 6 , 2 5 + 5 1 6 + 5 − 5 1 6 } . Thus, equality can be achieved.
So maximum value of P is S = 8 2 and ⌊ 1 0 0 S ⌋ = 9 0 5 .
Note: z , t could be simplitied to 2 5 ± 1 0 3 4 1 = 2 0 4 1 ± 3 .