In this paper, we consider the problem of modeling application
scenarios characterized by variability over time and involving
heterogeneous kinds of knowledge. The evolution of
distributed technologies creates new and challenging
possibilities of integrating different kinds of problem solving
methods, obtaining many benefits from the user point of view.
In particular, we propose here a multilayer modeling system
and adopt the Knowledge Artifact concept to tie together
statistical and Artificial Intelligence rule-based methods to
tackle problems in ubiquitous and distributed scenarios.