a + b 1 + a + c 1 + b + c 1 If a , b and c are non-negatve reals satisfying a b + b c + a c = 1 , find the minimum value of the expression above.
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.
Very nice!
I disagree with the reasoning presented in the first paragraph. Instead, we should simply say "Observe that f ( 1 , 1 , 0 ) = 2 5 . We will show that f ( a , b , c ) ≥ 2 5 , so this is indeed the minimium".
i solved it using the fact y x + x y > 2 , but i dont know if it is correct
Log in to reply
If you add your solution, the community can provide feedback on it.
Log in to reply
a + b + a + b 1 + b + c + b + c 1 + c + a + c + a 1 > 6
now we can get a + b + c > r o o t 3 so answer is 6-2.root3
Log in to reply
@Dev Sharma – Your solution (as currently stated) is wrong. Do you see why?
Note: It's also missing a lot of steps, so I don't know if you did think about it correctly.
Hint: Suppose that f ( a , b , c ) + g ( a , b , c ) > 6 and f ( a , b , c ) > 3 . What does that tell us about g ( a , b , c ) ?
Upon review, I found this solution wrong. I have failed to prove the minimum for positive reals.
a + b 1 + a + c 1 + b + c 1
= a c + b c c + a b + b c b + a b + a c a
= 1 − a b c + 1 − a c b + 1 − b c a
≥ 3 − ( a b + a c + b c ) ( a + b + c ) 2 = 2 ( a + b + c ) 2 by Cauchy-Schwarz inequality (T2' lemma)
≥ 2 9 ( a b c ) 3 1 by AM-GM inequality, with equality only holds when a = b = c = 3 1
Therefore, if a , b , c are all positive reals, the minimum is achived at 2 9 ( 3 1 ) = 2 3 3 > 2 5 .
Now consider one of a , b or c to be 0 . Assuming a = 0 , the expression becomes b 1 + c 1 + b + c 1 and we obtain b = c 1 from the constraint.
Writing everything in terms of b , we get:
b + b 1 + b + b 1 1 .
Now, applying the AM-GM inequality would yield a minimum of 2 . That sounds fantastic, until I realize that this requires b + b 1 = 1 , for which there is no real solution. Boo, what a letdown!
Now that the simple approach won't work, a little bashing using calculus seems like a nice idea...
Write f ( b ) = b + b 1 + b + b 1 1 = b ( b 2 + 1 ) ( b 2 + 1 ) 2 + b 2 . Compute the first derivative as f ′ ( b ) = b 2 ( b 2 + 1 ) 2 b 6 − 1 . Setting f ′ ( b ) = 0 , we find that there is only one turning point when b = 1 , lucky us! Computing the second derivative gives f ′ ′ ( 1 ) = 2 3 , so the point is indeed a minimum point and we can rest easy. Doing the same thing to f ( c ) yields the same result, so minimum is achieved when b = c = 1 . We check that the constraint b = c 1 is satisfied.
Therefore, when not all a , b , c are positive, the minimum is achieved when a = 0 , b = 1 , c = 1 or the its permutation. This gives a minimum of 2 5 .
The first half of the solution is incorrect. In particular, if we consider a path of points ( a n , b n , c n ) that approach ( 1 , 1 , 0 ) , then by the continuity of the functions, we should have lim f ( a n , b n , c n ) = 2 5 . For example, ( 0 . 9 , 0 . 9 , 1 . 8 0 . 1 9 ) is a counter example since the expression is equal to 2.544, which is less than 2 . 5 9 8 ≈ 2 3 3 . Do you see your mistake?
For the second half, we realize that we want to minimize f ( g ( b ) ) where g ( b ) = b + b 1 and f ( x ) + x + x 1 . Clearly, g ( b ) ≥ 2 , and thus from the shape of f ( x ) , we can conclude that the minimum of f ( g ( b ) ) is f ( 2 ) .
That's a nice solution, calculus did come in handy indeed
The first half of the solution is incorrect. In particular, if we consider a path of points ( a n , b n , c n ) that approach ( 1 , 1 , 0 ) , then by the continuity of the functions, we should have lim f ( a n , b n , c n ) = 2 5 . For example, ( 0 . 9 , 0 . 9 , 1 . 8 0 . 1 9 ) is a counter example since the expression is equal to 2.544, which is less than 2 . 5 9 8 ≈ 2 3 3 . Do you see your mistake?
For the second half, we realize that we want to minimize f ( g ( b ) ) where g ( b ) = b + b 1 and f ( x ) + x + x 1 . Clearly, g ( b ) ≥ 2 , and thus from the shape of f ( x ) , we can conclude that the minimum of f ( g ( b ) ) is f ( 2 ) .
Log in to reply
Yes, I realized my mistake soon after I submitted my solution.
Very nice analysis about how the second half can be proven.
Log in to reply
Can you identify the line where you made the mistake?
Log in to reply
@Calvin Lin – @Calvin Lin This part. I should have known earlier. Didn't know how can I make such obvious mistake while writing the proof.
The first line does not lead to the second.
≥ 2 9 ( a b c ) 3 1 by AM-GM inequality, with equality only holds when a = b = c = > 3 1
Therefore, if a , b , c are all positive reals, the minimum is achived at 2 9 ( 3 1 ) = 2 3 3 > 2 5 .
Log in to reply
@Zk Lin – Right, because we haven't bounded ( a b c ) as yet. For example, in the counterexample that I gave, 2 . 5 4 4 ≥ 2 9 ( 0 . 9 × 0 . 9 × 1 . 8 0 . 1 9 ) 3 1 . This lower bound becomes useless as c → 0 .
Problem Loading...
Note Loading...
Set Loading...
Call the expression P
By observation, we see that f ( 1 ; 1 ; 0 ) = 2 5 . So now we'll have to prove P ≥ 2 5 Without losing generality, assuming c ≤ b ≤ a , now we choose a number t > 0 satisfy a ≥ t ≥ b ≥ c and t 2 + 2 t c = a b + b c + c a = 1 ⇔ ( t + c ) 2 = ( a + c ) ( b + c ) = 1 + c 2
Now we prove a + b 1 + b + c 1 + a + c 1 ≥ t + c 2 + 2 t 1 ⇔ ( a + c 1 − b + c 1 ) 2 ≥ 2 t ( a + b ) ( a + c − b + c ) 2 ⇔ ( a + c ) ( b + c ) ≤ 2 t ( a + b ) which is right since a ≥ t ≥ b ≥ c
So now we need to prove t + c 2 + 2 t 1 ≥ 2 5 when a = b ≥ c .
From the condition of t we have c = 2 t 1 − t 2 ≥ 0 so we also find out that t ≤ 1
Subtitute and we get t 2 + 1 4 t + 2 t 1 ≥ 2 5 ⇔ 2 t ( t 2 + 1 ) − 5 t 3 + 9 t 2 − 5 t + 1 ≥ 0 Clearly see that the denominator is bigger than 0 so all we have left is − 5 t 3 + 9 t 2 − 5 t + 1 ≥ 0 ⇔ ( 1 − t ) ( 5 t 2 − 4 t + 1 ) ≥ 0 The above inequality is true because t ≤ 1 and 5 t 2 − 4 t + 1 > 0 so the equality hold when t = 1 ⇒ ( a ; b ; c ) = ( 1 ; 1 ; 0 ) and its permutations