3.1.5 Example 5

y=1y2y(0)=1

f(x,y)=1y2 is continuous in x everywhere. For y we want 1y20 or y21. The point y0=1 satisfies this. Now fy=2y21y2. We want 1y2>1 or y2<1. The point y0 does not satisfy this. Hence theory says nothing about uniqueness. Solution can be unique or not. When the ode has form y=f(y) we always check if IC satisfies the ode. In this case y(x)=1 does satisfy the ode. So this means y(x)=1 is solution. We do not need to solve by integration. But if we did, we will obtain the following

dy1y2=dxarcsin(y)=x+cy=sin(x+c)

At initial conditions the above gives 1=sinc. Hence c=π2. Therefore solution is y=sin(x+π2)=cosx. So this is another solution that satisfies the ode. Solution is not unique.