Difference inclusions arise naturally in the study of discrete-time or sampled-data systems.
We develop two novel sufficient conditions for robustness of a stability property referred to
as KL-stability with respect to an arbitrary measure; i.e., where a continuous positive
definite function of the solutions satisfies a class-KL estimate of time and the continuous
positive definite function of the initial condition.