Sum of Square Roots


By Graeme Mcrae on Thursday, January 03, 2002 - 12:02 am:

This question was given to me, which I thought would be fun for people to think about.


For non-negative numbers a, b, c and d,

a1

a+b5

a+b+c14

a+b+c+d30

Prove: a+b+c+d10

Please don't answer this if you've seen it (or one like it) before. I would like to leave the question on the board for a day or two so that people who haven't seen it before can try to work on it. Thanks!


By Michael Doré on Thursday, January 03, 2002 - 03:58 pm:

I think we can prove the more general result:

If x1 , ..., xn , y1 , ..., yn are non-negative and:

x1 y1

x1 + x2 y1 + y2

...

x1 ++ xn y1 ++ yn

then

x1 ++ xn y1 ++ yn

Then the result follows by setting n=4, x1 =a, x2 =b, x3 =c, x4 =d, y1 =1, y2 =4, y3 =9, y4 =16.


By Graeme Mcrae on Tuesday, January 08, 2002 - 04:56 am: A good way to solve this puzzle is to make a substitution. If we define new variables represented by capital A, B, C, and D as follows, then the solution is easier:

Let A=a-1

Let B=b-2

Let C=c-3

Let D=d-4

Now, solving for a, b, c, and d, we get the following equations.

a= A2 +2A+1

b= B2 +4B+4

c= C2 +6C+9

d= D2 +8D+16

Since a1, and a+b5, etc., we get the following four inequalities:

a= A2 +2A+11

a+b= A2 + B2 +2A+4B+55

a+b+c= A2 + B2 + C2 +2A+4B+6C+1414

a+b+c+d= A2 + B2 + C2 + D2 +2A+4B+6C+8D+3030

Notice that in each inequality the same constant appears on both sides, so they cancel:

A2 +2A0

A2 + B2 +2A+4B0

A2 + B2 + C2 +2A+4B+6C0

A2 + B2 + C2 + D2 +2A+4B+6C+8D0

A square is always non-negative, so it can be subtracted from the "less than" side of an equality:

2A0

2A+4B0

2A+4B+6C0

2A+4B+6C+8D0

Now multiply the first inequality by 6, the second by 2, the third by 1, and the fourth by 3:

6(2A)0

2(2A+4B)0

1(2A+4B+6C)0

3(2A+4B+6C+8D)0

Now add up the four equations, then divide by 24:

24A+24B+24C+24D0

A+B+C+D0

Now substitute the lower-case expressions in place of A, B, C, and D, and then add 10 to both sides:

a-1+b-2+c-3+d-40

a+b+c+d10


That's it! Pretty easy, don't you think?


By Kerwin Hui on Tuesday, January 08, 2002 - 06:07 pm:
Michael's conjecture should also have the condition yi increasing. It is very easy to find a counterexample otherwise (e.g. a4, a+b5, with a=2.25, b=2.56, gives a+b=3.1>4+1). The proof is very similar to the solution of the problem.

Kerwin