We first show that f is injective. Suppose that f(y)=f(z). Then we have:
x+yxy=f(f(x)+f(y))=f(f(x)+f(z))=x+zxz
Taking reciprocals of the first and last expression implies y=z, as required for injectivity.
Note that the right hand side of the functional equation can be written as
x+yxy=x1+y11
If x=n−ak and y=n+ak with a=±n, the expression x1+y1=k2n does not depend on a. Hence, for such x and y, the right hand side of the functional equation has a value that does not depend on a. In particular, with n≥2, a=0 and a=1, the functional equation implies
f(f(nk)+f(nk))=f(f(n−1k)+f(n+1k)).
As f is injective, we can equate the arguments of the two outer f's to deduce:
f(n+1k)−f(nk)=f(nk)−f(n−1k).(20)
Next, we are going to prove that there are rational numbers a,b such that
f(x)=a+xbfor all x∈Q+.(21)
Our first proof of (21) starts with the claim that for each k∈Q+ and any integer n≥1 there exist rational numbers a(k) and b(k) such that
f(nk)=a(k)+b(k)kn.(22)
Indeed, from (20) we know that
b(k):=k(f(n+1k)−f(nk))
does not depend on n. In particular, we have kb(k)=f(2k)−f(k). If we let a(k)=2f(k)−f(2k) we obtain f(k)=a(k)+kb(k). This is the case n=1 of (20) which gets our proof by induction going. For the inductive step, we use the definition of b(k) and the inductive hypothesis:
f(n+1k)=f(nk)+kb(k)=a(k)+b(k)kn+kb(k)=a(k)+b(k)kn+1.
We next show that a(k) and b(k) do not depend on k. For any k∈Q+ and n∈Z+ we have
f(nk)=f(1k′)andf(2nk)=f(2k′)wherek′=nk
and (22) gives us
a(k)+b(k)kn=a(k′)+b(k′)knanda(k)+b(k)k2n=a(k′)+b(k′)k2n.
Subtracting the second from twice the first equation, and the first from the second, gives
a(k)=a(k′)andb(k)kn=b(k′)k′nhence
a(k)=a(nk)andb(k)=b(nk)for all k∈Q+,n∈Z+.
Letting k=1 we now obtain a(1)=a(n1) for all n∈Z+. Using k=n/m with arbitrary m,n∈Z+ we obtain a(mn)=a(m1). Together these show that a(k)=a(1) for all k∈Q+. Similarly it follows that b(k)=b(1) for all k∈Q+. This establishes (21) with a=a(1) and b=b(1).
For our second proof of (21) we define
g(k):=f(n+1k)−f(nk)
For all k∈Q+ and integers n≥1. From (20) we know that the right hand side does not depend on n. A straightforward induction yields
f(n+ik)=f(nk)+ig(k)for all k∈Q+,n≥1,i≥0.(23)
Next we show that vg(1)=ug(u/v) for all positive integers u,v. To see this, we first substitute k=1 and i=n=v in equation (23) to obtain
f(2v1)=f(v1)+vg(1).
Next we substitute k=u/v and i=n=u in equation (23) and obtain
f(2v1)=f(v1)+ug(vu).
Comparing these two equations we get the desired equality, which means that kg(k)=g(1) for all k∈Q+.
Note that substituting n=1 and i=m−1≥0 in (23) gives
f(mk)=f(k)+(m−1)g(k).(24)
For k∈Q+ we now define a(k)=f(k)−g(k). Setting k=1 in (24) gives
f(m1)=f(1)+(m−1)g(1)=a(1)+mg(1).
With k=n/m and m replaced by n in (24) we obtain
f(m1)=f(nmn)=f(mn)+(n−1)g(mn)=a(mn)+ng(mn)=a(mn)+mg(1).
Comparing these two equations, we obtain a(n/m)=a(1) for all positive integers m,n. If we now let a=a(1) and b=g(1) we finally obtain
f(k)=a(k)+g(k)=a(1)+kg(1)=a+kb
for all k∈Q+. This finishes the second proof of (21).
Our penultimate step is to show that a=0 in (21). To show this, we start by substituting (21) into the original functional equation with y=x:
2x=f(f(x)+f(x))=f(2f(x))=f(2a+x2b)=a+2ax+2bbx.
This implies x(ax+b)=2a(ax+b)+bx, i.e. a(2ax−x2+2b)=0. So, either a=0 or 2b=x2−2ax for all x∈Q+. But the polynomial x2−2ax is not constant, so we must have a=0. Thus, f(x)=xb with b∈Q+. An easy check reveals that all such functions satisfy the functional equation. As f(1)=2023, we must have b=2023 so that f(2023)=1.