Processing math: 100%

Monday, December 21, 2015

Prove a(ac)2+b(bc)2(ac)(bc)(a+bc) for all real a,b,c0: First Attempt

For reals a,b,c0, prove the inequality

a(ac)2+b(bc)2(ac)(bc)(a+bc)

and state when the equality holds.

It's hard not to relate the LHS of the inequality with the AM-GM inequality formula, but if we persist, we would apply the AM-GM inequality to the LHS, and we get:

a(ac)2+b(bc)22a(ac)2b(bc)22(ac)(bc)ab

If we're heading in the right direction, we would definitely get the following inequality that holds for all a,b,c0.

2(ac)(bc)ab(ac)(bc)(a+bc)

The above is true if a=b=c=0 or a=b=c0.

If abc, we need to prove 2ab(a+bc) also holds for all positive real a,b,c.

2(ac)(bc)ab(ac)(bc)(a+bc)

2ab(a+bc)

Squaring both sides of the inequality:

4ab(a+b)22c(a+b)+c2

4aba2+2ab+b22ac2bc+c2

2ab+2bc+2aca2+b2+c2

Replace 2ab+2bc+2ac by 2ab+2bc+2ac=(a+b+c)2(a2+b2+c2), the above inequality becomes

(a+b+c)2(a2+b2+c2)a2+b2+c2

(a+b+c)22(a2+b2+c2)---(*)

But from the Cauchy–Schwarz inequality, we have:

a2+b2+c2a2+b2+c2ab+bc+ac

a2+b2+c2ab+bc+ac

2(a2+b2+c2)2(ab+bc+ac)

2(a2+b2+c2)+a2+b2+c2a2+b2+c2+2(ab+bc+ac)

3(a2+b2+c2)(a+b+c)2---(**)

The result from (**) doesn't suggest (*) must be correct.

Thus we're back to square one.

I encourage you to try the problem and I'll lead you to the proof on my next blog post.

No comments:

Post a Comment