= State transition matrix
{wiki}
For a homogeneous system $x'=A(t)x$, its <state transition matrix> $\Phi(t,s)$ maps the state at time $s$ to the state at time $t$:
$$
\partial_t\Phi(t,s)=A(t)\Phi(t,s),\quad\Phi(s,s)=I,\quad x(t)=\Phi(t,s)x(s).
$$
For continuous coefficients on an interval, uniqueness gives $\Phi(t,r)\Phi(r,s)=\Phi(t,s)$ and $\Phi(t,s)^{-1}=\Phi(s,t)$. If $X$ is any <fundamental matrix>, then $\Phi(t,s)=X(t)X(s)^{-1}$, independent of that choice. A forced system has the <variation of constants> formula
$$
x(t)=\Phi(t,s)x(s)+\int_s^t\Phi(t,r)f(r)\,dr.
$$
For constant $A$, $\Phi(t,s)=e^{(t-s)A}$. Generally it is not $\exp(\int_s^tA)$, since matrices at different times need not commute. For piecewise constant $A=A_1$ on $[0,1)$ and $A=A_2$ on $[1,2]$, chronological propagation gives $\Phi(2,0)=e^{A_2}e^{A_1}$. With $A_1=\begin{pmatrix}0&1\\0&0\end{pmatrix}$ and $A_2=\begin{pmatrix}0&0\\1&0\end{pmatrix}$, this product is $\begin{pmatrix}1&1\\1&2\end{pmatrix}$, while $e^{A_1+A_2}$ has both diagonal entries $\cosh1$. This exhibits the importance of time order.<Ordinary Differential Equations and Dynamical Systems by Gerald Teschl>[§3.4, general linear first-order systems]{ref}
Back to article page