Maths Olympiad Prep

Library / /5 of 5

, 2014

Algebra Difficulty 9.1 IMO level Prove it Germany

Determine the smallest positive integer nn for which there exists a polynomial
P(X)=a2nX2n+a2n1X2n1++a1X+a0 P(X)=a_{2 n} X^{2 n}+a_{2 n-1} X^{2 n-1}+\ldots+a_{1} X+a_{0}
with real coefficients that has both of the following properties:
- For i=0,1,,2ni=0,1, \ldots, 2 n we have 2014ai20152014 \leq a_{i} \leq 2015.
- There exists a real number ξ\xi with P(ξ)=0P(\xi)=0.

Solution

We have n=2014n=2014. First, for a special choice of the coefficients aka_{k}, namely 2014 for even kk and 2015 for odd kk, define the polynomial Qn(X)=2014X2n+2015X2n1+2014X2n2++2015X+2014Q_{n}(X)=2014 X^{2 n}+2015 X^{2 n-1}+2014 X^{2 n-2}+\ldots+2015 X+2014. We have Q2014(1)=0Q_{2014}(-1)=0, so the number sought is at most 2014. It remains to show that P(X)P(X) has no real zero for n2013n \leq 2013. For x0x \geq 0 we have P(x)a0>0P(x) \geq a_{0}>0, since the coefficients are positive; for x<0x<0 we have P(x)Qn(x)P(x) \geq Q_{n}(x), since xkx^{k} is positive for even kk, hence akxk2014xka_{k} x^{k} \geq 2014 x^{k}, and xkx^{k} is negative for odd kk, hence akxk2015xka_{k} x^{k} \geq 2015 x^{k}. It remains to show that Qm(x)>0Q_{m}(x)>0 for m2013m \leq 2013.

1st proof: One computes that for real xx we have
Qm(x)=1007(ν=0m1(xν+xν+1)2)+12(ν=0m1(x2ν+1+1)(x2m2ν1+1))+(1007m2)(x2m+1) Q_{m}(x)=1007\left(\sum_{\nu=0}^{m-1}\left(x^{\nu}+x^{\nu+1}\right)^{2}\right)+\frac{1}{2}\left(\sum_{\nu=0}^{m-1}\left(x^{2 \nu+1}+1\right)\left(x^{2 m-2 \nu-1}+1\right)\right)+\left(1007-\frac{m}{2}\right)\left(x^{2 m}+1\right)
The summands of the first sum are non-negative, being squares. Since x2k+1+1x^{2 k+1}+1 is positive, 0, or negative, respectively, for all k0k \geq 0 for x>1x>-1, x=1x=-1, or x<1x<-1, respectively, the summands in the second sum are non-negative. Thus for x<0x<0 and m2013m \leq 2013 we have:
P(x)Qm(x)(1007m2)(x2m+1)>12 P(x) \geq Q_{m}(x) \geq\left(1007-\frac{m}{2}\right)\left(x^{2 m}+1\right)>\frac{1}{2}

2nd proof (using analysis): For x<0x<0 define the real function fm(x)=(x21)Qm(x)=2014(x2m+21)+2015x(x2m1)f_{m}(x)=\left(x^{2}-1\right) Q_{m}(x)=2014\left(x^{2 m+2}-1\right)+2015 x\left(x^{2 m}-1\right). The second derivative fm(x)=2(2m+1)(2014(m+1)x2m+2015mx2m1)f_{m}^{\prime \prime}(x)=2(2 m+1)\left(2014(m+1) x^{2 m}+2015 m x^{2 m-1}\right) has the unique negative zero xm=2015m2014(m+1)>1x_{m}=-\frac{2015 m}{2014(m+1)}>-1 for m2013m \leq 2013, and we have
fm(xm)=2014(2m+2)xm2m+1+2015(2m+1)xm2m2015==(2014(2m+2)xm+2015(2m+1))xm2m2015=2015xm2m2015<0 \begin{aligned} f_{m}^{\prime}\left(x_{m}\right) & =2014(2 m+2) x_{m}^{2 m+1}+2015(2 m+1) x_{m}^{2 m}-2015= \\ & =\left(2014(2 m+2) x_{m}+2015(2 m+1)\right) x_{m}^{2 m}-2015=2015 \cdot x_{m}^{2 m}-2015<0 \end{aligned}
For x<xmx<x_{m} or xm<x<0x_{m}<x<0, respectively, fmf_{m}^{\prime} is strictly increasing or strictly decreasing, respectively, since the derivative fmf_{m}^{\prime \prime} is positive or negative, respectively; hence fmf_{m}^{\prime} has a global maximum at x=xmx=x_{m} for x<0x<0 and is negative for all x<0x<0. Hence fmf_{m} is strictly decreasing, and thus, because of fm(1)=0f_{m}(-1)=0, is positive for x>1x>-1 and negative for x<1x<-1. Thus Qm(x)Q_{m}(x) is positive for x1x \neq-1; for x=1x=-1 we directly get Qm(1)=2014m>0Q_{m}(-1)=2014-m>0.

Want a route through all this instead of an archive? The track puts 2,444 problems in a working order, from Junior Challenge level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement translated into English from de; metadata (topic, difficulty) added by this project.