×

# RMO 2014 Delhi Region Q.6

Let $$x_1,x_2,\ldots ,x_{2014}$$ be positive real numbers such that $\large\displaystyle\sum^{2014}_{j=1} x_j=1$ Determine with proof the smallest constant $$K$$ such that $\large K \displaystyle\sum^{2014}_{j=1} \dfrac{x^2_j}{1-x_j}\geq1$

• You can find rest of the problems here

• You can find the solutions here

Note by Aneesh Kundu
2 years, 5 months ago

Sort by:

Method 1

$$x_{1} + x_{2} + \ldots + x_{2014} = 1$$

$$\dfrac{x_{1}^{2}}{1 - x_{1}^{2}} + \dfrac{x_{2}^{2}}{1 - x_{2}^{2}} + ........................... + \dfrac{x_{201}4^{2}}{1 - x_{2014}^{2}}$$

Applying Titu's Lemma ,

$$\dfrac{x_{1}^{2}}{1 - x_{1}^{2}} + \dfrac{x_{2}^{2}}{1 - x_{2}^{2}} + ........................... + \dfrac{x_{201}4^{2}}{1 - x_{2014}^{2}} \geq \dfrac{(x_{1} + x_{2} + .... + x_{2014})^{2}}{2014 - (x_{1} + x_{2} + .....+ x_{2014})}$$

$$\dfrac{x_{1}^{2}}{1 - x_{1}^{2}} + \dfrac{x_{2}^{2}}{1 - x_{2}^{2}} + ........................... + \dfrac{x_{201}4^{2}}{1 - x_{2014}^{2}} \geq \frac{1}{2013}$$

Thus , $$\boxed{K = 2013}$$

Method 2

$$\displaystyle \sum_{1}^{2014} \dfrac{x^{2}_{j} - 1}{1 - x_{j}} + \dfrac{1}{1 - x_{j}}$$

$$\displaystyle \sum_{1}^{2014} -( 1 + x_{j}) + \dfrac{1}{1 - x_{j}}$$

$$\boxed{- 2015 + \displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}} = \displaystyle \sum_{1}^{2014} \dfrac{x^{2}_{j}}{1 - x_{j}}}$$

Applying $$A.M \geq H.M$$

$$\frac{\displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}}}{2014} \geq \dfrac{2014}{ 2014 - ( x_{1} + x_{2} ..... x_{2014})}$$

$$\displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}} \geq \dfrac{2014^{2}}{ 2013}$$

$$- 2015 + \displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}} \geq \dfrac{2014^{2}}{ 2013} - 2015$$

$$- 2015 + \displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}} \geq \dfrac{2014^{2}- 2015 \times 2013}{2013}$$

$$- 2015 + \displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}} \geq \dfrac{2014^{2} - (2014 + 1)(2014 - 1)}{2013}$$

$$- 2015 + \displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}} \geq \dfrac{2014^{2} - (2014^{2} - 1)}{2013}$$

$$- 2015 + \displaystyle \sum_{1}^{2014} \dfrac{1}{1 - x_{j}} \geq \dfrac{1}{2013}$$

$$\displaystyle \sum_{1}^{2014} \dfrac{x^{2}_{j}}{1 - x_{j}} \geq \dfrac{1}{2013}$$

$$2013\displaystyle \sum_{1}^{2014} \dfrac{x^{2}_{j}}{1 - x_{j}} \geq 1$$

$$\boxed{K = 2013}$$

Enjoy!!!

$\Huge\color{#ff00b3}{♛}\;\;\;\color{#ff0000}{ ❤ }\;\;\;\color{#0000ff}{\mathbf{B}}\color{#ff7f00}{\mathbf{r}}\color{#ffff00}{\mathbf{i}}\color{#00ff00}{\mathbf{l}}\color{#00ffff}{\mathbf{l}}\color{#0000ff}{\mathbf{i}}\color{#8b00ff}{\mathbf{a}}\color{#ff0000}{\mathbf{n}}\color{#ff7f00}{\mathbf{t}}\color{color:#ffff00}{\mathbf{.}}\color{#00ff00}{\mathbf{o}}\color{#0081ff}{\mathbf{r}}\color{#ff69b4}{\mathbf{g}}$ copied from anastisya romoniva comment

· 2 years, 5 months ago

Your second solution using AM>HM is what I wrote during the RMO, almost verbatim... Wow. · 2 years, 5 months ago

Yes i too enjoyed very much doing by the 2nd method · 2 years, 5 months ago

u can use \dfrac(it gives bigger fractions) and also \ldots(it gives the dots) · 2 years, 5 months ago

I applied the \Idots , its not working · 2 years, 5 months ago

Its actually small "L" $$\ldots$$

I know its confusing sometimes · 2 years, 5 months ago

Most latex is small letters, unless you want to stress something or make it

More lines: \rightarrow, \RIghtarrow give us $$\rightarrow , \Rightarrow$$
Capital Greek alphabet: \gamma, \Gamma give us $$\gamma, \Gamma$$ Staff · 2 years, 5 months ago

Yeah!! · 2 years, 5 months ago

Its tagged under A.M G.M inequality , have you done using it , if yes then how · 2 years, 5 months ago

Actually I meant AM-HM

Typo · 2 years, 5 months ago

Contrary to @Sandeep Rathod solution there is one more method :

Let an function $$y\quad =\quad \cfrac { \quad { x }^{ 2 } }{ 1-x } \\$$. in (0,1)

So it inscribe an convex polygon of centroid G so from jensons inequality :

$$y\quad =\quad f(x)\quad =\quad \cfrac { \quad { x }^{ 2 } }{ 1-x } \\ \\ { y }_{ G }\quad \ge \quad { y }_{ p }\\ \\ \frac { \sum { f\left( { x }_{ i } \right) } }{ n } \quad \ge \quad f\left( \cfrac { \sum { { x }_{ i } } }{ n } \right) \\ \\ \sum { f\left( { x }_{ i } \right) } \quad \ge \quad nf\left( \cfrac { 1 }{ n } \right) \quad \quad \quad (put\quad n\quad =\quad 2014)\\ \\ 2013\sum { f\left( { x }_{ i } \right) } \quad \ge \quad 1\\ \\ \boxed { K\quad =\quad 2013 }$$.

Q.E.D · 2 years, 5 months ago

i did'nt understood , how it's inscribed? @Deepanshu Gupta

I appreciate - you always think different · 2 years, 5 months ago

Thanks ! But Your Solution is also very elegant ! I also appreciate it !

And Actually if we consider 'n' points (2014 points) on curve and join them, then they will form an closed loop (or we can say convex polygon ) So Centroid of this Polygon , which is surely lies inside the polygon So it's y-coordinate is greater or equal (when all pt's are collinear , I think) to y-coordinates of point on curve which has same x-coordinate as that of centroid of this loop (Polygon) !

Here I used Jenson's inequality · 2 years, 5 months ago

I learnt the spelling of "Cauchy Schwartz" today,and I got the solution at my first glance! ;-) · 2 years, 5 months ago

I think there's no 't'. · 2 years, 5 months ago

Yeah! You are right. I got the spelling wrong! * sob * @Joel Tan · 2 years, 5 months ago

Please post your method @Ranjana Kasangeri · 2 years, 5 months ago

Sir,Multiply the LHS by summation or (1-x) ,then apply Cauchy! · 2 years, 5 months ago

I got 2013 by Cauchy-Scwartz · 2 years, 5 months ago

Ditto. Do you think solving 4 is enough!Dunno, this waz my only RMO. · 2 years, 5 months ago

I used Titu's Lemma for this one! · 2 years, 5 months ago

I had given my RMO yesterday and done this using Jensen's inequality. · 2 years, 5 months ago

Chebyshev · 2 years, 5 months ago

Tchebycheff · 2 years, 5 months ago

-- http://en.wikipedia.org/wiki/Chebyshev%27sinequality · 2 years, 5 months ago

i got $$2013$$ · 2 years, 5 months ago

Anish can you please post the first 5 questions also... This one I did by cauchy Schwartz · 2 years, 5 months ago

Its given below the question · 2 years, 5 months ago

Can someone explain the third question please · 2 years, 5 months ago

Is that the one about permuting the digits of two powers of 2? · 2 years, 5 months ago