Half-trivial isn't that trivial!

Algebra Level 5

Let a , b , c a,b,c be non-negative reals so that a + b + c = 3 a+b+c=3 . Find the sum of the maximum value and the minimum value of a + a b + a b c a+ab+abc . Write your answer to 2 decimal places.

Source : AoPS community.


The answer is 4.0.

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.

3 solutions

Nguyễn Anh
Aug 11, 2017

Let S = a + a b + a b c S = a + ab + abc

As the other solution has mentioned, it is apparent that the minimum value of S S is 0 0 since a , b , c 0 a, b, c \geq 0

I have another approach to find the maximum of S S

S = a ( 1 + b + b c ) S = a(1 + b + bc)

Substituting a = 3 b c a = 3 - b - c

S = ( 3 b c ) ( 1 + b + b c ) S = (3 - b - c)(1 + b + bc)

S = 3 b c + 3 b b 2 b c + 3 b c b 2 c b c 2 \Leftrightarrow S = 3 - b - c + 3b - b^{2} - bc + 3bc - b^{2}c - bc^{2}

S = c 2 b c ( b 2 2 b + 1 ) ( b 2 2 b + 1 ) + 4 \Leftrightarrow S = -c^{2}b - c(b^{2} - 2b + 1) - (b^{2} - 2b + 1) + 4

S = [ c 2 b + ( c + 1 ) ( b 1 ) 2 ] + 4 \Leftrightarrow S = -[c^{2}b + (c+1)(b-1)^{2}] + 4

Since c 2 b 0 c^{2}b \geq 0 and ( c + 1 ) ( b 1 ) 2 0 (c+1)(b-1)^{2} \geq 0

S 4 \Rightarrow S \leq 4

The inequality happens when c 2 b = ( c + 1 ) ( b 1 ) 2 = 0 c^{2}b = (c+1)(b-1)^{2} = 0 , or a = 2 , b = 1 , c = 0 a = 2, b = 1, c = 0

\Rightarrow Sum of the maximum and minimum value of S = a + a b + a b c S = a + ab + abc is 4 4

This is a better solution. I fixed one value say a. Tried to find the maximum value assuming a=k

Srikanth Tupurani - 2 years, 3 months ago

Most of the Vietnamese try to avoid calculus and think something totally different. This is a much elegant solution.

Srikanth Tupurani - 2 years, 3 months ago
Marco Brezzi
Jul 29, 2017

Relevant wiki: Lagrange Multipliers

Since a , b , c 0 a,b,c \geq 0 , a + a b + a b c a+ab+abc can't be negative and its minimum value is 0 0 (that occurs when a = 0 a=0 )

Now we are going to search the maximum value using the Lagrange multipliers method:

L ( a , b , c , λ ) = a + a b + a b c λ ( a + b + c 3 ) \mathcal{L}(a,b,c,\lambda)=a+ab+abc-\lambda(a+b+c-3)

Taking the partial derivatives with respect to a a , b b and c c and setting them equal to 0 0 we get

{ 1 + b + b c = λ a + a c = λ a b = λ \begin{cases} 1+b+bc=\lambda \\ a+ac=\lambda \\ ab=\lambda \end{cases}

Substituting λ = a b \lambda=ab in the second equation and simplifying a a (that can't be 0 0 since that would give us the minimum)

a + a c = a b b = c + 1 a+ac=ab \iff b=c+1

Substituting b = c + 1 b=c+1 in the first equation

1 + c + 1 + ( c + 1 ) c = λ ( c + 1 ) 2 + 1 = λ 1+c+1+(c+1)c=\lambda \iff (c+1)^2+1=\lambda

Since λ = a ( c + 1 ) \lambda=a(c+1) from the second equation, substituting it in the first we get a quadratic in c + 1 c+1

( c + 1 ) 2 a ( c + 1 ) + 1 = 0 (c+1)^2-a(c+1)+1=0

Solving for c + 1 c+1

c + 1 = a ± a 2 4 2 2 c = a ± a 2 4 2 c+1=\dfrac{a\pm\sqrt{a^2-4}}{2} \iff 2c=a\pm\sqrt{a^2-4}-2

Substituting b = c + 1 b=c+1 and 2 c = a ± a 2 4 2 2c=a\pm\sqrt{a^2-4}-2 into the original constraint a + b + c = 3 a+b+c=3

a + c + 1 + c = 3 a + 2 c = 2 2 a ± a 2 4 = 4 \begin{aligned} & a+c+1+c=3\\ \iff \quad & a+2c=2\\ \iff \quad & 2a\pm\sqrt{a^2-4}=4 \end{aligned}

Solving separately the + + and - cases we get from both a = 2 a=2 (the - case also gives a = 10 3 a=\dfrac{10}{3} , but we have to discard it because it would make b b or c c negative), which implies c = 2 ± 2 2 4 2 2 = 0 c=\dfrac{2\pm\sqrt{2^2-4}-2}{2}=0 and b = 0 + 1 = 1 b=0+1=1 , therefore the maximum value of a + a b + a b c a+ab+abc is 2 + 2 1 + 2 1 0 = 4 2+2\cdot 1+2\cdot 1\cdot 0=4

In conclusion the sum of the maximum and minimum values of a + a b + a b c a+ab+abc is

4 + 0 = 4 4+0=\boxed{4}

Note: By right, you have to add in the conditions of a 0 , b 0 , c 0 a \geq 0, b \geq 0 , c \geq 0 in the initial Lagrangian.

Your calculation is for the relative extrema based only on a + b + c = 3 a+b+c = 3 , and a , b , c a, b, c are any real values. When you reject the case of a = 10 3 a = \frac{ 10}{3} because "it would make b b or c c negative", you do not know if there is another relative minimum on the restricted domain.

E.g. If we wanted to minimize x 2 + 2 x + 1 x^2 + 2x + 1 on the domain x 0 x \geq 0 , we cannot say that "Well, the relative minimum on the unrestricted domain occurs at x = 1 x = -1 . Let's ignore it. Hence, there is no relative minimum on the restricted domain."


This is why calculus proofs of inequalities at the Olympiad are either completely correct or scored at 0. There is almost no partial credit given.

Calvin Lin Staff - 3 years, 10 months ago

Log in to reply

What would be a better way to treat that a = 10 3 a=\dfrac{10}{3} solution? Also could you please check this problem which I think is wrong? (I also posted a report )

(Sorry, I don't know how to write hyperlinks from my mobile device)

Marco Brezzi - 3 years, 10 months ago

Log in to reply

You would have to use the correct domain in the Lagrangian, and apply the boundary conditions. IE You will need to check what happens at a = 0 , b = 0 , c = 0 , a = 0 , b =0, c=0, etc. Think about how this relates to the case of x 2 + 2 x + 2 x^2 + 2x + 2 on x 0 x \geq 0 . What do we need to do?


I see that you filed a report on that problem. Give the problem creator time to respond to it. I check back in a week (usu over the weekend) on these reports if there is no response.


To hyperlink, we use Markdown syntax of [display text](URL) . I have edited your comment.

Calvin Lin Staff - 3 years, 10 months ago

I love this solution! Interesting use of partial derivatives that I never would have thought of :)

Zach Abueg - 3 years, 10 months ago

Log in to reply

Check out the lagrange multipliers wiki. This is a "standard calculus" technique, and it's one of the reasons why inequalities have fallen out of favor in Olympiads.

Calvin Lin Staff - 3 years, 10 months ago

Log in to reply

I'm on it, thanks! Haha, it seems like the prevailing attitude in Olympiad circles is that calculus is unrigorous and doesn't get to the essence of the problem.

Zach Abueg - 3 years, 10 months ago

Log in to reply

@Zach Abueg No. Calculus is rigorous. However, there are a lot of potential misconceptions (in part because it is not taught rigorously in high school) E.g. Can you spot an issue with the above proof? If not, see my above conversation with Marco.

Olympiads do not reward unrigorous/sloppy proofs. It doesn't matter which "style" they are in. Other examples

  • In geometry, coordinate geometry is often graded on a "complete or 0" scale, though you could get partial credit for stating and proving important observations.
  • In inequalities, algebraic manipulation/bashing is also graded on a "complete or 0" scale.

Calvin Lin Staff - 3 years, 10 months ago

Log in to reply

@Calvin Lin Yes, calculus techniques are often used very badly, which earns no sympathy. I personally don't have any knowledge of Lagrange multipliers, but from what I've heard from friends, when used correctly, they can really hack an inequality problem without revealing anything about its underlying structure. Take problem 2 from IMO 2002, for instance: there are solutions that involve Lagrange multipliers, and also those where it is factored as f ( x , y , z ) 2 0 f(x, y, z)^2 \geq 0 , the latter of which seems to have been intended by problem writers to better get to the heart of the problem.

Zach Abueg - 3 years, 10 months ago
Frank Petiprin
Sep 26, 2017
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
#Pythonista Numerical Solution for various interval divisions on [0,3] 
#Using a triple loop with index variables a, b, c
increment=.03125 #.0625#.125 #0.25 Four Runs Of Program
minnum=100000
maxnum=0
findmm=0
loop=int((3/increment)+1)
for a in range(0,loop): 
    ar=a*increment
    for b in range(0,loop):
      br=b*increment
      for c in range(0,loop):
         cr=c*increment
         sumabcr=ar+br+cr
         if(sumabcr==3):
           findmm=ar+(ar*br)+(ar*br*cr)
           if(findmm<minnum):
               minnum=findmm
           if(findmm>maxnum):
               maxnum=findmm
print('Final Min =',minnum,' Max =',maxnum)
print(' For Loop Increment',increment)
#Comment: Sample Runs! It works with other increments such as 0.1,1/3

Final Min = 0.0  Max = 4.0
 For Loop Increment 0.25
++++++++++++++++++++++++
++++++++++++++++++-+++++++
Final Min = 0.0  Max = 4.0
 For Loop Increment 0.125
++++++++++++++++++++++++
++++++++++++++++++-+++++++
Final Min = 0.0  Max = 4.0
 For Loop Increment 0.0625
++++++++++++++++++++++++
++++++++++++++++++-+++++++
Final Min = 0.0  Max = 4.0
 For Loop Increment 0.03125
++++++++++++++++++++++++ 

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...