2.20.3.4.1.1 Algorithm Given \(y^{\prime }=f_{0}+f_{1}y+f_{2}y^{2}\) let \(y=u\left ( x\right ) e^{\phi }\) where \(\phi =\int f_{1}dx\) and the ode becomes
Where \(F\left ( x\right ) =f_{0}e^{-\phi },G\left ( x\right ) =f_{2}e^{\phi }\). There is one special subcase here to consider. If \(F\left ( x\right ) \) turns out to be proportional to \(G\left ( x\right ) \) then the ode is separable. So this check should be done after completing the above. See second example below of one such example.
2.20.3.4.1.2 Example \(y^{\prime }=x+3y+7y^{2}\) Comparing to \(y^{\prime }=f_{0}+f_{1}y+f_{2}y^{2}\) shows that
Let
Therefore \(\phi =\int 3dx=3x\) and we have
We see that the transformed \(u\) ode will always have the general form
Where
Hence the linear term \(f_{1}\) is removed. Now we will try to solve (1). It is not one of the special cases we have solved before. We can either try to find a particular solution, or if we can’t, then convert it to second order ode and solve it that way.
2.20.3.4.1.3 Example \(y^{\prime }=5xe^{3x}+3y+10xe^{-3x}y^{2}\) Comparing to \(y^{\prime }=f_{0}+f_{1}y+f_{2}y^{2}\) shows that
Let
Therefore \(\phi =\int 3dx=3x\) and we have
Hence
Since \(F\left ( x\right ) \) is proportional to \(G\left ( x\right ) \), then this is separable ode which is easily solved. Once \(u\) is known, then \(y\) is found since \(y=u\left ( x\right ) e^{\phi }\).