Olympiad Maths Prep

Library / /27 of 29

Algebra Difficulty 7.3 National olympiad, round 2 Prove it Iran

We are given a natural number dd. Find all open intervals of maximum length IRI \subseteq \mathbb{R} such that for all real numbers a0,a1,,a2d1a_0, a_1, \dots, a_{2d-1} inside the interval II, the polynomial P(x)=x2d+a2d1x2d1++a1x+a0P(x) = x^{2d} + a_{2d-1}x^{2d-1} + \dots + a_1x + a_0 has no real roots.

Solution

The answer is I=(1,1+1d)I = (1, 1 + \frac{1}{d}).

Assume that the desired interval is of the form (b,c)(b, c). For some q,rq, r in (b,c)(b, c), put ai=qa_i = q for odd and ai=ra_i = r for even ii, where 0i2d10 \le i \le 2d - 1. Then
P(1)=1dq+dr. P(-1) = 1 - dq + dr.
Since P(x)P(x) has no real root, we must have P(1)>0P(-1) > 0. Thus, qr<1dq - r < \frac{1}{d}. We can assume that cb=1dc - b = \frac{1}{d}. Therefore, consider the interval as (b,b+1d)(b, b + \frac{1}{d}). It is easy to find that b>0b > 0, since by choosing a0=b+ϵa_0 = b + \epsilon should be positive. Now, assign the following numbers to the coefficients of the polynomial.
a2d1=a2d3==a1=b+1d+ϵ,a2d=a2d2==a0=b+ϵ, a_{2d-1} = a_{2d-3} = \dots = a_1 = b + \frac{1}{d} + \epsilon, \quad a_{2d} = a_{2d-2} = \dots = a_0 = b + \epsilon,
for some sufficiently small ϵ>0\epsilon > 0.
It is clear that for all positive real xx, P(x)>0P(x) > 0. For all negative real xx, putting x=tx = -t, where t>0t > 0, then,
P(t)=t2d(b+1d)t2d1+bt2d2(b+1d)t+b+ϵQ(t), P(-t) = t^{2d} - \left(b + \frac{1}{d}\right) t^{2d-1} + bt^{2d-2} - \dots - \left(b + \frac{1}{d}\right) t + b + \epsilon Q(t),
for some polynomial Q(t)Q(t), degQ(t)=2d1\deg Q(t) = 2d - 1. As ϵ\epsilon tends to zero, it remains to find all b>0b > 0 such that
R(t)=t2d(b+1d)t2d1+bt2d2(b+1d)t+b0. R(t) = t^{2d} - \left(b + \frac{1}{d}\right) t^{2d-1} + b t^{2d-2} - \cdots - \left(b + \frac{1}{d}\right) t + b \geq 0.
Note that R(1)=0R(1) = 0. Therefore, R(1)R'(1) should be zero. That is,
2d(2d1)(b+1d)+(2d2)b(b+1d)=0. 2d - (2d - 1) \left(b + \frac{1}{d}\right) + (2d - 2)b - \cdots - \left(b + \frac{1}{d}\right) = 0.
Hence, we claim the interval I=(1,1+1d)I = (1, 1 + \frac{1}{d}) works! It is obvious that P(x)P(x) has no positive real roots. Now, we prove that P(x)>0P(x) > 0 for all negative real xx. Put x=t,t>0x = -t, t > 0.
Then,
P(t)>t2d(1+1d)t2d1+t2d2(1+1d)t+1. P(-t) > t^{2d} - \left(1 + \frac{1}{d}\right) t^{2d-1} + t^{2d-2} - \dots - \left(1 + \frac{1}{d}\right) t + 1.
Thus it remains to prove that
t2d+t2d2++t2+1d+1t2d1+t2d3++t3+td. \frac{t^{2d} + t^{2d-2} + \dots + t^2 + 1}{d+1} \ge \frac{t^{2d-1} + t^{2d-3} + \dots + t^3 + t}{d}.
Now, since t2d+1t2d2k+1+t2k1t^{2d} + 1 \geq t^{2d-2k+1} + t^{2k-1} and t2k+t2k22t2k1t^{2k} + t^{2k-2} \geq 2t^{2k-1}, we are done.
Thus,
b1dlimt1+dt2d(t2d1+t2d3++t)t2d1t2d2++t1. b \le \frac{1}{d} \lim_{t \to 1^+} \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t - 1}.
and
b1dlimt1dt2d(t2d1+t2d3++t)t2d1t2d2++t1, b \geq \frac{1}{d} \lim_{t \to 1^-} \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t - 1},
since the function
dt2d(t2d1+t2d3++t)t2d1t2d2++t1 \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t - 1}
has a limit as tt approaches 1. Thus, we find that
limt1+dt2d(t2d1+t2d3++t)t2d1t2d2++t1=limt1dt2d(t2d1+t2d3++t)t2d1t2d2++t1. \lim_{t \to 1^+} \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t-1} = \lim_{t \to 1^-} \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t-1}.
Hence,
b=1dlimt1dt2d(t2d1+t2d3++t)t2d1t2d2++t1. b = \frac{1}{d} \lim_{t \to 1} \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t - 1}.
Finally note that
dt2d(t2d1+t2d3++t)t2d1t2d2++t1=(t2dt2d1)+(t2dt2d3)++(t2dt)(t1)(t2d2+t2d4++1)=(t1)(t2d1+t2d3(t2+t+1)++t(t2d2++1))(t1)(t2d2+t2d4++1) \begin{align*} & \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t-1} \\ &= \frac{(t^{2d} - t^{2d-1}) + (t^{2d} - t^{2d-3}) + \dots + (t^{2d} - t)}{(t-1)(t^{2d-2} + t^{2d-4} + \dots + 1)} \\ &= \frac{(t-1)(t^{2d-1} + t^{2d-3}(t^2+t+1) + \dots + t(t^{2d-2} + \dots + 1))}{(t-1)(t^{2d-2} + t^{2d-4} + \dots + 1)} \end{align*}
Thus,
limt1dt2d(t2d1+t2d3++t)t2d1t2d2++t1 \lim_{t \to 1} \frac{dt^{2d} - (t^{2d-1} + t^{2d-3} + \dots + t)}{t^{2d-1} - t^{2d-2} + \dots + t - 1}
=limt1t2d1+t2d3(t2+t+1)++t(t2d2++1)t2d2+t2d4++1=d2d=1. = \lim_{t \to 1} \frac{t^{2d-1} + t^{2d-3} (t^2 + t + 1) + \dots + t (t^{2d-2} + \dots + 1)}{t^{2d-2} + t^{2d-4} + \dots + 1} = \frac{d^2}{d} = 1.
That is, b=1db = \frac{1}{d} and d=1d = 1.

Looking for a route rather than an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement and solution reproduced as published; topic and difficulty added by this site.