Concept:
For any general Linear Time-Invariant (LTI) system, stability is classified under the Bounded-Input Bounded-Output (BIBO) standard. A system is defined as BIBO stable if every bounded input produces a bounded output. The necessary and sufficient criterion to guarantee this in the time domain relates directly to the absolute integrability of its impulse response $h(t)$.
Step 1: Derivation of the stability criterion.
Let the input signal to an LTI system be $x(t)$, bounded by a finite real maximum value $M_x$:
\[
|x(t)| \le M_x < \infty \quad \forall \quad t
\]
The output $y(t)$ of the system is given by the convolution integral of the input and the impulse response:
\[
y(t) = \int_{-\infty}^{\infty} h(\tau) x(t - \tau) \, d\tau
\]
Taking the absolute value of both sides:
\[
|y(t)| = \left| \int_{-\infty}^{\infty} h(\tau) x(t - \tau) \, d\tau \right|
\]
Applying the triangle inequality for integrals, the absolute value of an integral is less than or equal to the integral of the absolute value:
\[
|y(t)| \le \int_{-\infty}^{\infty} |h(\tau)| \cdot |x(t - \tau)| \, d\tau
\]
Since $|x(t-\tau)| \le M_x$, we can substitute this upper bound into the inequality:
\[
|y(t)| \le M_x \int_{-\infty}^{\infty} |h(\tau)| \, d\tau
\]
Step 2: Imposing the absolute integrability constraint.
For the output $y(t)$ to remain strictly bounded ($|y(t)| < \infty$), given that $M_x$ is a finite non-zero constant, the remaining integral factor must evaluate to a finite value:
\[
\int_{-\infty}^{\infty} |h(\tau)| \, d\tau < \infty
\]
This mathematically demonstrates that the impulse response $h(t)$ must be absolutely integrable. Additionally, causality implies $h(t) = 0$ for $t < 0$, which restricts the limits from $0$ to $\infty$, but absolute integrability remains the core prerequisite for stability.