The part that is giving me trouble is:
$$\mathcal{L}[xy']$$
I have never done this before so I ploughed through.
$$\mathcal{L}[xy'] = \int^{\infty}_0ty'(t)e^{-pt}dt = -\frac{d}{dp}\int^{\infty}_0y'(t)e^{-pt}dt = -\frac{d}{dp}\mathcal{L}[y'] = \frac{d}{dp}[y(0) - \mathcal{L}[y]]$$
so in the grand scheme of things $$\frac{d}{dp}[y(0) - \mathcal{L}[y]] + \mathcal{L}[y] = \frac{k!}{p^{k+1}}$$
How can I proceed?
Could I have done something differently to get something nicer?
$$\color{red}{\text{The restriction that I use a Laplace transform is necessary. }}$$
After a few trial and errors I came up with this;
If we make the substitution $x = e^z$
then $$xy'=x\frac{dy}{dx} = e^z\frac{dy}{dz}(\frac{dy}{dx})^{-1}=\frac{dy}{dz}$$
so the equation in question becomes:
$$y' + y = e^{kz}$$
on taking the Laplace transform of both sides
$$p\mathcal{L}[y]-y_0 + \mathcal{L}[y] = \frac{1}{p-k}$$
$$\implies \mathcal{L}[y] = \frac{1+y_0(p-k)}{(p-k)(p+1)}$$
$$\implies \mathcal{L}[y] = \frac{1}{(p-k)(p+1)}+\frac{y_0}{(p+1)}$$
$$\implies y = y_0e^{-z} + \frac{e^{-z}+e^{kz}}{1+k}$$
On making our substitutions
$$y = \frac{y_0}{x}+\frac{\frac{1}{x}+x^k}{k+1}$$
It would be great if someone could refute/confirm the validity of my solution.
It does satisfy the differential equation.