Let d=D(a,b). We can write a=a1d and b=b1d, where a1 and b1 are co-prime. We have v(a,b)=a1b1d. We insert this in the condition of the problem to get
a1d+b1d∣d+a1b1d⇒a1+b1∣1+a1b1,
and we can rearrange the inequality to
a1d+b1dd+a1b1d≤4a1d+b1d⇔1+a1b1≤d4(a1+b1)2⇔4≤d(a12+b12)+(2d−4)a1b1.
If d≥2 we have
d(a12+b12)+(2d−4)a1b1≥2(a12+b12)+(2⋅2−4)a1b1=2(a12+b12)≥2(12+12)=4,
Lets now check the case when d=1. We must prove the inequality
4≤a12+b12−2a1b1⇔4≤(a1−b1)2⇔2≤∣a1−b1∣.
This means we have to prove that the numbers a1 and b1 differ by at least 2. Hence it suffices to check if that they do not differ for less.
If a1=b1 we get a1=b1=1 due to co-primality and hence a=b=1, which is impossible.
If the numbers differ by 1 we may assume that b1=a1+1 since the other case is symmetric. We insert this to the condition of the problem to get
a1+(a1+1)∣1+a1(a1+1)⇒2a1+1∣2(a12+a1+1)⇒2a1+1∣a1+2=a12+a1+1=(2a1+1)a1+a1+2
However, since 2a1+1≥a1+2 this is possible only if a1=1. Since d=1 we again get a=1, which is impossible since a,b>1.
Thus we proved that the numbers a1 and b1 differ by at least 2 which proves the inequality.