First, note that ⌊kn⌋ represents the number of multiples of k among 1,2,…,n. Hence,
k=1∑n⌊kn⌋τ(k)2=k=1∑n1≤m≤n, k∣m∑τ(k)2=m=1∑nk∣m∑τ(k)2.
To prove the problem statement, it suffices to justify, for m=1,…,n,
5ω(m)≤k∣m∑τ(k)2≤5Ω(m).
When m=1, it is obvious. When m>1, let m=p1α1p2α2⋯prαr be the prime factorization (p1,p2,…,pr are distinct prime numbers and α1,α2,…,αr are positive integers). For k=p1β1p2β2⋯prβr, τ(k)=(β1+1)(β2+1)⋯(βr+1). Thus,
k∣m∑τ(k)2=0≤β1≤α10≤βr⋯≤αr∑(β1+1)2(β2+1)2⋯(βr+1)2=i=1∏r(12+22+⋯+(αi+1)2).
Now it suffices to show for each 1≤i≤r,
5≤12+22+⋯+(αi+1)2≤5αi.
For j∈N+, define T(j)=12+22+⋯+(j+1)2=61(j+1)(j+2)(2j+3).
Then
T(j)T(j+1)=j+1j+2⋅j+2j+3⋅2j+32j+5∈[1,12⋅23⋅35]=[1,5].
Since T(1)=5, it is straightforward to check T(j)∈[5,5j] (∀j∈N+) by induction. This finishes the proof. □