Stability in Continuous-Time LTI Difference Equation Model w/ No Input
For an input-lesscontinuous-time LTI modeldtdx(t)=Ax(t) with x(0)=x0, where A∈Rn×n has eigenvalues λ1,...,λn, the model is:
Internally stable, if ℜ{λi}<0 for all i.
Marginally stable, if
ℜ{λi}≤0
There exists at least one λm where ℜ{λm}=0
All poles with a zero real part must be unique.
Continuous-time stability can be illustrated on the Cartesian plane as
\usepackage{tikz}\usepackage{pgfplots}\begin{document}\begin{tikzpicture} % draw the coordinates \draw[->] (-2.5, 0) -- (2.5, 0) node[right] {$Re\{\lambda\}$}; \draw[->] (0,-2.5) -- (0,2.5) node[above] {$Im\{\lambda\}$}; \draw[->, thick, color=yellow] (0,-2.5) -- (0,2.5); \pgfsetfillopacity{0.25} % draw the unit circle \draw[thick, color=red!0, fill=red!30] (0,2.5)--(-2.5, 2.5)--(-2.5,-2.5)--(0,-2.5);\end{tikzpicture}\end{document}
Where the light-red-filled area (Re{λ}<1) is where internal stability is guaranteed, and the yellow line (Re{λ}=1) is where marginal stability may occur. Non-colored areas (Re{λ}>0) indicate where internal stability is guaranteed not to occur.
Intuition
Let’s first try to see why internal stability only happens in the above scenarios. Consider the input-less dxdxc(t)=Acxc(t). The solution for this