2f(y2)=f(y)2.(4)
Using this, we can cancel the 2f(y2) on the left hand side of the original functional equation against the f(y)2 on the right hand side:
f(x2y)=x2f(y).
Substituting y=1 in this equation yields f(x2)=x2f(1), and substituting y=−1 yields f(−x2)=x2f(−1). Because x2 takes on all non-negative numbers as value when x∈R, we get
f(x)={cxdxals x≥0,als x<0,
with c=f(1) and d=−f(−1). Now substitute y=1 in equation (4), which yields 2f(1)=f(1)2, hence 2c=c2. It follows that c=0 or c=2. If we actually substitute y=−1 in equation (4), then we find that 2f(1)=f(−1)2, hence 2c=(−d)2. For c=0, we get d=0, and for c=2, we get d=2 or d=−2. So, there are three cases:
* c=0,d=0: then f(x)=0 for all x;
* c=2,d=2: then f(x)=2x for all x;
* c=2,d=−2: then f(x)=2x for x≥0, and f(x)=−2x for x<0, or, in other words, f(x)=2∣x∣ for all x.
Using the first function, both sides of the functional equation become 0, so this function is a solution. Using the second function, we get 2x2y+4y2 on the left hand side, and (x2+2y)⋅2y=2x2y+4y2 on the right hand side, so this function is a solution as well. Using the third function, we get 2∣x2y∣+4∣y2∣=2x2∣y∣+4y2 on the left hand side, and (x2+2∣y∣)⋅2∣y∣=2x2∣y∣+4∣y∣2=2x2∣y∣+4y2 on the right hand side, so also this function is a solution.
Altogether, we found the three solutions: f(x)=0, f(x)=2x, and f(x)=2∣x∣. □