For any n>1, we can write the given equality for n and n−1, then subtract them side by side to obtain the following:
i=1∑n−1(a⌊in⌋−a⌊in−1⌋)+a1=n10−(n−1)10.
Defining a new sequence by b1=a1=1 and bn=an−an−1 for each n>1, we can express the equality we obtained in the following form:
i∣n∑bin=n10−(n−1)10⇒i∣n∑bi=n10−(n−1)10.
It is easy to see that this equality holds for n=1 as well as n>1. Now, recall the definition of the well-known μ-function:
μ(n)={(−1)k0n is square-free and the number of prime divisors of n is kn is not square-free
Also recall the following well-known theorem:
Theorem. Let x1,x2,… be a sequence and let yn=∑i∣nxi for each positive integer n. Then,
xn=i∣n∑μ(i)⋅yin
holds for each positive integer n. If n>1, denoting the prime divisors of n by p1,p2,…,pk, we can express the equality more explicitly as follows:
xn=yn−yp1n−⋯−ypkn+yp1p2n+⋯+ypk−1pkn−…
Now, we know that
i∣n∑bi=n10−(n−1)10=(10n9−45n8+⋯−45n2+10n−1).
Therefore, by the theorem above, we find that
bn=i∣n∑μ(i)⋅((in)10−(in−1)10)
=10⋅i∣n∑μ(i)⋅(in)9−45⋅i∣n∑μ(i)⋅(in)8+⋯+10⋅i∣n∑μ(i)⋅(in)−i∣n∑μ(i).
For n>1, denoting the prime divisors of n by p1,p2,…,pk, we see that the following holds for each positive integer r:
i∣n∑μ(i)⋅(in)r=nr⋅(1−p1r1)⋯(1−pkr1)=ϕ(n)⋅nr−1⋅(1+p11+⋯+p1r−11)⋯(1+pk1+⋯+pkr−11).
Furthermore, ∑i∣nμ(i)=0 since n>1, therefore we obtain the following:
ϕ(n)bn=10n8∏(1+p1+⋯+p81)−45n7∏(1+p1+⋯+p71)+⋯−45n∏(1+p1)+10.
We draw the following two conclusions from this formula:
(i) For each positive integer n, one has ϕ(n)∣bn. This is obvious for n=1 and follows from the formula for n>1.
(ii) For each positive integer n, one has bn>0, thus an≥n. Because,
b1=1>0, b2=210−2>0, b3=310−210−1>0, b4=410−310−210+1>0
and for n≥5,
ϕ(n)bn>n7⋅(10n−45)∏(1+p1+⋯+p71)+…+n⋅(120n−45)∏(1+p1)+10>0⇒bn>0.
For n>1, separate the prime divisors of n into two subsets as those which divide c and those which do not divide c, hence write n=m⋅s. Here, m divides sufficiently large powers of c and s is coprime to c. Hence, m∣cn−1∣can−1 since an−1≥n−1 and s∣cϕ(s)−1∣cϕ(n)−1∣cbn−1 since ϕ(n)∣bn. Therefore, n∣can−1⋅(cbn−1)=can−can−1.