Maths Olympiad Prep

Library / /313 of 397

Algebra Difficulty 6.7 National Olympiad Prove it Taiwan

Let a0,a1,a2,a_0, a_1, a_2, \dots be a sequence of real numbers such that a0=0,a1=1a_0 = 0, a_1 = 1, and for every n2n \ge 2 there exists 1kn1 \le k \le n satisfying
an=an1++ankk. a_n = \frac{a_{n-1} + \cdots + a_{n-k}}{k}.
Find the maximum possible value of a2018a2017a_{2018} - a_{2017}.

Solutions — 2

Solution 1

The claimed maximal value is achieved at
a1=a2==a2016=1,a2017=a2016++a02017=112017, a_1 = a_2 = \cdots = a_{2016} = 1, \quad a_{2017} = \frac{a_{2016} + \cdots + a_0}{2017} = 1 - \frac{1}{2017},
a2018=a2017++a12017=1120172. a_{2018} = \frac{a_{2017} + \cdots + a_1}{2017} = 1 - \frac{1}{2017^2}.

Now we need to show that this value is optimal. For brevity, we use the notation
S(n,k)=an1+an2++ank for nonnegative integers kn. S(n, k) = a_{n-1} + a_{n-2} + \cdots + a_{n-k} \text{ for nonnegative integers } k \le n.
In particular, S(n,0)=0S(n, 0) = 0 and S(n,1)=an1S(n, 1) = a_{n-1}. In these terms, for every integer n2n \ge 2 there exists a positive integer knk \le n such that an=S(n,k)/ka_n = S(n, k)/k.
For every integer n1n \ge 1 we define
Mn=max1knS(n,k)k,mn=min1knS(n,k)k,andΔn=Mnmn0. M_n = \max_{1 \le k \le n} \frac{S(n, k)}{k}, \quad m_n = \min_{1 \le k \le n} \frac{S(n, k)}{k}, \quad \text{and} \quad \Delta_n = M_n - m_n \ge 0.
By definition, an[mn,Mn]a_n \in [m_n, M_n] for all n2n \ge 2; on the other hand,
an1=S(n,1)1[mn,Mn]. a_{n-1} = \frac{S(n, 1)}{1} \in [m_n, M_n].
Therefore,
a2018a2017M2018m2018=Δ2018, a_{2018} - a_{2017} \le M_{2018} - m_{2018} = \Delta_{2018},

Also by definition, for any 0<kn0 < k \le n we have
kmnS(n,k)kMn; km_n \le S(n, k) \le kM_n;
notice that these inequalities are also valid for k=0k = 0.

Claim 1. For every n>2n > 2, we have Δnn1nΔn1\Delta_n \le \frac{n-1}{n}\Delta_{n-1}.

Proof. Choose positive integers k,nk, \ell \le n such that Mn=S(n,k)/kM_n = S(n, k)/k and mn=S(n,)/m_n = S(n, \ell)/\ell. We have
S(n,k)=an1+S(n1,k1), S(n, k) = a_{n-1} + S(n-1, k-1),
so
k(Mnan1)=S(n,k)kan1=S(n1,k1)(k1)an1(k1)(Mn1an1), \begin{aligned} k(M_n - a_{n-1}) &= S(n, k) - ka_{n-1} = S(n-1, k-1) - (k-1)a_{n-1} \\ &\le (k-1)(M_{n-1} - a_{n-1}), \end{aligned}
since
S(n1,k1)(k1)Mn1. S(n-1, k-1) \le (k-1)M_{n-1}.
Similarly, we get
(an1mn)=(1)an1S(n1,1)(1)(an1mn1). \ell(a_{n-1} - m_n) = (\ell - 1)a_{n-1} - S(n-1, \ell-1) \le (\ell - 1)(a_{n-1} - m_{n-1}).
Since mn1an1Mn1m_{n-1} \le a_{n-1} \le M_{n-1} and k,nk, \ell \le n, the obtained inequalities yield
Mnan1k1k(Mn1an1)n1n(Mn1an1)andan1mn1(an1mn1)n1n(an1mn1). \begin{aligned} M_n - a_{n-1} &\le \frac{k-1}{k}(M_{n-1} - a_{n-1}) \le \frac{n-1}{n}(M_{n-1} - a_{n-1}) \quad \text{and} \\ a_{n-1} - m_n &\le \frac{\ell-1}{\ell}(a_{n-1} - m_{n-1}) \le \frac{n-1}{n}(a_{n-1} - m_{n-1}). \end{aligned}

Δn=(Mnan1)+(an1mn)n1n((Mn1an1)+(an1mn1))=n1nΔn1.\begin{align*} \Delta_n &= (M_n - a_{n-1}) + (a_{n-1} - m_n) \\ &\le \frac{n-1}{n} \left( (M_{n-1} - a_{n-1}) + (a_{n-1} - m_{n-1}) \right) \\ &= \frac{n-1}{n} \Delta_{n-1}. \end{align*}

Back to the problem, if an=1a_n = 1 for all n2017n \le 2017, then a20181a_{2018} \le 1 and hence
a2018a20170.a_{2018} - a_{2017} \le 0.
Otherwise, let 2q20172 \le q \le 2017 be the minimal index with aq<1a_q < 1. We have S(q,i)=iS(q, i) = i for all i=1,2,,q1i = 1, 2, \dots, q-1, while S(q,q)=q1S(q, q) = q-1. Therefore, aq<1a_q < 1 yields
aq=S(q,q)q=11q. a_q = \frac{S(q, q)}{q} = 1 - \frac{1}{q}.
Now we have
S(q+1,i)=i1q for i=1,2,,q, S(q + 1, i) = i - \frac{1}{q} \text{ for } i = 1, 2, \dots, q,
and
S(q+1,q+1)=q1q. S(q + 1, q + 1) = q - \frac{1}{q}.
This gives us
mq+1=S(q+1,1)1=S(q+1,q+1)q+1=q1qand m_{q+1} = \frac{S(q+1, 1)}{1} = \frac{S(q+1, q+1)}{q+1} = \frac{q-1}{q} \quad \text{and}
Mq+1=S(q+1,q)q=q21q2, M_{q+1} = \frac{S(q+1, q)}{q} = \frac{q^2 - 1}{q^2},
so
Δq+1=Mq+1mq+1=(q1)q2. \Delta_{q+1} = M_{q+1} - m_{q+1} = \frac{(q-1)}{q^2}.

Solution 2

We present a different proof of the estimate a2018a2017201620172a_{2018} - a_{2017} \le \frac{2016}{2017^2}.
We keep the same notations of S(n,k)S(n, k), mkm_k and MnM_n from the previous solution.

Notice that S(n,n)=S(n,n1)S(n, n) = S(n, n - 1), as a0=0a_0 = 0. Also notice that for 0kln0 \le k \le l \le n we have
S(n,l)=S(n,k)+S(nk,lk). S(n, l) = S(n, k) + S(n - k, l - k).

Proof. Choose a positive integer kn+1k \le n + 1 such that
mn+1=S(n+1,k)k. m_{n+1} = \frac{S(n+1, k)}{k}.
Then we have
kmn+1=S(n+1,k)=an+S(n,k1)mn+(k1)mn=kmn, km_{n+1} = S(n+1, k) = a_n + S(n, k-1) \ge m_n + (k-1)m_n = km_n,
which establishes the first inequality in the Claim. The proof of the second inequality is similar.

Proof. By Claim 2, we have
[mk,Mk][mk1,Mk1][mn,Mn]. [m_k, M_k] \subseteq [m_{k-1}, M_{k-1}] \subseteq \dots \subseteq [m_n, M_n].
Since ak[mk,Mk]a_k \in [m_k, M_k], the claim follows.

Proof. We use induction on nn. The case n=2n = 2 is routine. To perform the induction step, we need to prove the inequalities
S(n,n)nS(n,k)kandS(n,k)kS(n,n1)n1(1) \frac{S(n, n)}{n} \le \frac{S(n, k)}{k} \quad \text{and} \quad \frac{S(n, k)}{k} \le \frac{S(n, n-1)}{n-1} \qquad (1)
for every positive integer knk \le n. Clearly, these inequalities hold for k=nk = n and k=n1k = n - 1, as S(n,n)=S(n,n1)>0S(n, n) = S(n, n - 1) > 0. In the sequel, we assume that k<n1k < n - 1.
Now the first inequality in Eq. (1) rewrites as
nS(n,k)kS(n,n)=k(S(n,k)+S(nk,nk)), nS(n, k) \ge kS(n, n) = k(S(n, k) + S(n - k, n - k)),
or, cancelling the terms occurring on both parts, as
(nk)S(n,k)kS(nk,nk)S(n,k)kS(nk,nk)nk. (n-k)S(n,k) \ge kS(n-k, n-k) \Leftrightarrow S(n,k) \ge k \cdot \frac{S(n-k, n-k)}{n-k}.
By the induction hypothesis, we have S(nk,nk)/(nk)=mnkS(n-k, n-k)/(n-k) = m_{n-k}.
By Claim 3, we get animnka_{n-i} \ge m_{n-k} for all i=1,2,,ki = 1, 2, \dots, k. Summing these kk inequalities we obtain
S(n,k)kmnk=kS(nk,nk)nk, S(n, k) \ge k m_{n-k} = k \cdot \frac{S(n-k, n-k)}{n-k},
as required.

The second inequality in Eq. (1) is proved similarly. Indeed, this inequality is equivalent to
(n1)S(n,k)kS(n,n1)(nk1)S(n,k)kS(nk,nk1)S(n,k)kS(nk,nk1)nk1=kMnk; \begin{align*} (n-1)S(n,k) &\le kS(n,n-1) \\ \Leftrightarrow (n-k-1)S(n,k) &\le kS(n-k,n-k-1) \\ \Leftrightarrow S(n,k) &\le k \cdot \frac{S(n-k,n-k-1)}{n-k-1} = kM_{n-k}; \end{align*}
the last inequality follows again from Claim 3, as each term in S(n,k)S(n, k) is at most MnkM_{n-k}.

Now we can prove the required estimate for a2018a2017a_{2018} - a_{2017}. Set N=2017N = 2017. By Claim 4,
aN+1aNMN+1aN=S(N+1,N)NaN=aN+S(N,N1)NaN=S(N,N1)NN1NaN. \begin{align*} a_{N+1} - a_N \le M_{N+1} - a_N &= \frac{S(N+1, N)}{N} - a_N \\ &= \frac{a_N + S(N, N-1)}{N} - a_N \\ &= \frac{S(N, N-1)}{N} - \frac{N-1}{N} \cdot a_N. \end{align*}
On the other hand, the same Claim yields
aNmN=S(N,N)N=S(N,N1)N. a_N \ge m_N = \frac{S(N, N)}{N} = \frac{S(N, N-1)}{N}.
Noticing that each term in S(N,N1)S(N, N-1) is at most 1, so
S(N,N1)N1, S(N, N-1) \le N-1,
we finally obtain
aN+1aNS(N,N1)NN1NS(N,N1)N=S(N,N1)N2N1N2. \begin{align*} a_{N+1} - a_N &\le \frac{S(N, N-1)}{N} - \frac{N-1}{N} \cdot \frac{S(N, N-1)}{N} \\ &= \frac{S(N, N-1)}{N^2} \le \frac{N-1}{N^2}. \end{align*}

Want a route through all this instead of 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 translated into English from the original; metadata (topic, difficulty) added by this project.