Titu's lemma

From testwiki
Revision as of 21:03, 14 August 2024 by imported>David Eppstein (Undid revision 1240306772 by LR.127 (talk) encyclopedia articles are generally not about terms; see WP:NOTDICT. If you do not understand what this article is actually about, then maybe you should not be the one trying to summarize it in a short description.)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Template:Short description Template:Multiple issues In mathematics, the following inequality is known as Titu's lemma, Bergström's inequality, Engel's form or Sedrakyan's inequality, respectively, referring to the article About the applications of one useful inequality of Nairi Sedrakyan published in 1997,[1] to the book Problem-solving strategies of Arthur Engel published in 1998 and to the book Mathematical Olympiad Treasures of Titu Andreescu published in 2003.[2][3] It is a direct consequence of Cauchy–Bunyakovsky–Schwarz inequality. Nevertheless, in his article (1997) Sedrakyan has noticed that written in this form this inequality can be used as a proof technique and it has very useful new applications. In the book Algebraic Inequalities (Sedrakyan) several generalizations of this inequality are provided.[4]

Statement of the inequality

For any real numbers a1,a2,a3,,an and positive reals b1,b2,b3,,bn, we have a12b1+a22b2++an2bn(a1+a2++an)2b1+b2++bn. (Nairi Sedrakyan (1997), Arthur Engel (1998), Titu Andreescu (2003))

Probabilistic statement

Similarly to the Cauchy–Schwarz inequality, one can generalize Sedrakyan's inequality to random variables. In this formulation let X be a real random variable, and let Y be a positive random variable. X and Y need not be independent, but we assume E[|X|] and E[Y] are both defined. Then E[X2/Y]E[|X|]2/E[Y]E[X]2/E[Y].

Direct applications

Example 1. Nesbitt's inequality.

For positive real numbers a,b,c: ab+c+ba+c+ca+b32.

Example 2. International Mathematical Olympiad (IMO) 1995.

For positive real numbers a,b,c, where abc=1 we have that 1a3(b+c)+1b3(a+c)+1c3(a+b)32.

Example 3.

For positive real numbers a,b we have that 8(a4+b4)(a+b)4.

Example 4.

For positive real numbers a,b,c we have that 1a+b+1b+c+1a+c92(a+b+c).

Proofs

Example 1.

Proof: Use n=3, (a1,a2,a3):=(a,b,c), and (b1,b2,b3):=(a(b+c),b(c+a),c(a+b)) to conclude: a2a(b+c)+b2b(c+a)+c2c(a+b)(a+b+c)2a(b+c)+b(c+a)+c(a+b)=a2+b2+c2+2(ab+bc+ca)2(ab+bc+ca)=a2+b2+c22(ab+bc+ca)+112(1)+1=32.

Example 2.

We have that (1a)2a(b+c)+(1b)2b(a+c)+(1c)2c(a+b)(1a+1b+1c)22(ab+bc+ac)=ab+bc+ac2a2b2c23a2b2c232a2b2c2=32.

Example 3.

We have a21+b21(a+b)22 so that a4+b4=(a2)21+(b2)21(a2+b2)22((a+b)22)22=(a+b)48.

Example 4.

We have that 1a+b+1b+c+1a+c(1+1+1)22(a+b+c)=92(a+b+c).

References

Template:Reflist

Template:Linear algebra