Let k be a positive integer. Show that if there exists a sequence a0,a1,… of integers satisfying the condition an=nan−1+nk for all n≥1, then k−2 is divisible by 3.
Proposed by Okan Tekman, Turkey
This one wants a proof. Work it on paper, then read the official solution and mark
yourself. Be honest about it: the record is only any use to you if it is.
Official solution
1. Lemma 1: There exists a polynomial P(x)∈Z[x] with degree k−1 such that xP(x)−P(x−1)=xk+c for some constant c∈Z.
Proof: Write P(x)=ak−1xk−1+ak−2xk−2+⋯+a1x+a0. Now, solve for the ai's in descending order, starting with ak−1 and working downwards. By comparing the xk-coefficients on both sides of the equation, we clearly have ak−1=1. If ai+1,⋯,ak−1 have been determined for some i≥0, then ai is uniquely determined by examining the xi+1-coefficients on both sides of the equation. Furthermore, we can see that ai is integral. ■
2. As guaranteed by Lemma 1, let P(x) be a monic polynomial of degree k−1 with integer coefficients and c∈Z be an integer such that xP(x)−P(x−1)=xk+c.
3. Lemma 2: c=0 and an=P(n) for all n∈Z≥0.
Proof: Let bn=P(n)−an for all n∈Z≥0. Note that from the condition of the problem: nan−an−1=nk,∀n∈N. From the equation xP(x)−P(x−1)=xk+c, we have: nP(n)−P(n−1)=nk+c,∀n∈N. Therefore, subtracting the two equations gives: nbn−bn−1=c,∀n∈N. Let M=max(∣b0∣,∣c∣). We claim now that this implies bi∈[−2M,2M] for all i∈Z≥0. For i=0,1, this is immediate from the definition of M. For i>1, we can prove it easily by induction on i. Hence, we've shown that the sequence {bi} is bounded by 2M in absolute value.
4. Now, notice that for n>3M, we have: ∣bn∣=nbn−1+c≤nbn−1+nc<3M2M+3MM=1, and hence bn=0. Therefore, c=(3M+2)b3M+2−b3M+1=0, as claimed. From here, it easily follows from downwards induction (with base case n=3M+1) that bn=0 for all n∈Z≥0. This implies that P(n)−an=0⇒an=P(n) for all n∈Z≥0. ■
5. From the lemma, as c=0, we know that: xP(x)−P(x−1)=xk.
6. Let's take this equation modulo x4+x2+1. It's clear that the remainder of P modulo x4+x2+1 is of the form R(x)=ax3+bx2+cx+d for some a,b,c,d∈Z. Now, observe that P(x)−R(x)=(x2+x+1)(x2−x+1)Q(x) for some Q∈Z[x]. In particular, observe that x2−x+1∣P(x)−R(x) and x2−x+1∣P(x−1)−R(x−1), since (x−1)2+(x−1)+1=x2−x+1. Therefore, viewing the equation in modulo x2−x+1, we obtain: xR(x)−R(x−1)≡xk.
7. Plugging the formula for R into the equation yields: ax4+bx3+cx2+dx−a(x−1)3−b(x−1)2−c(x−1)−d≡xk. Noting that for i∈Z≥2, we have xi≡xi−1−xi−2 and (x−1)i≡−(x−1)i−1−(x−1)i−2, we have that: (c+d−a)x−(b+c)+(b−c)x+(c−a−d)≡xk, i.e. (b+d−a)x−(a+b+d)≡xk.
8. Observe that: 1≡0⋅x+1, x≡1⋅x+0, x3≡−1, and x4≡(−1)⋅x+0. So since b+d−a,a+b+d are either both 0 or both nonzero (since a,b,d∈Z), we cannot have k∈{0,1,3,4}. Furthermore, as xk≡xk+6 for k∈Z≥0, we actually have that k cannot be 0,1,3, or 4 (mod 6). In other words, k≡2 (mod 3), as desired. ■
The final answer is k−2 is divisible by 3.
Source: NuminaMath-1.5,
licensed Apache-2.0.
Statement and solution reproduced as published; topic, difficulty and ordering added
by this site.