I have an exercise where i have to find the general solution to this problem:
$$ X'=\left( \begin{matrix} 2&-1\\ 4&-2 \end{matrix} \right)X + \begin{pmatrix} 2\\1 \end {pmatrix}. $$ Obviously the system matrix has linearly dependent vectors, which means that it can't be diagonalsed. Another method I tried is to use an expansion and using it I got a homogeneous solution:
$$ X_h = e^{tA} \begin{pmatrix}C_1\\C_2\end{pmatrix} $$ Where $$ e^{tA} = I + tA + \frac{t^2A^2}{2!} + ... , where A^2 = 0 $$ The homogeneous solution is then $$ \begin{pmatrix} 1+2t&-t\\ 4t&1-2 \end{pmatrix} * \begin{pmatrix} C_1\\C_2 \end{pmatrix} $$ The problem is I dont know how to get the particular solution. I tried using $$ X_p = (sI - A)^{-1}\pmatrix{2\\1}e^{st} $$ Where s in this case is zero, but the matrix -A is not invertible. Can anyone help me?
If $A^2=0$ you can find easily $e^{At}=I+tA$ by the Taylor expansion. Now the solution is given by the formula $$ X(t)=e^{At}X(0)+e^{At}\int_0^te^{-As}\left[\matrix{2\\1}\right]\,ds.\tag1 $$ P.S. To get $(1)$ use the integrating factor $e^{-At}$ for the system $$ \underbrace{e^{-At}(X'-AX)}_{(e^{-At}X(t))'}=e^{-At}\left[\matrix{2\\1}\right]\quad\Leftrightarrow\quad e^{-At}X(t)-X(0)=\int_0^t e^{-As}\left[\matrix{2\\1}\right]\,ds. $$
P.P.S. How the integrating factor works