A Model-Based Systems Synthesis Approach to Set and Solve a Deployment Problem of Software Functions on a Heterogeneous Embedded Hardware Architecture
Stephen Creff, Pierre‐Alain Yvars, Laurent Zimmer, Jérôme Le Noir, Sébastien Madelénat · IEEE Transactions on Systems Man and Cybernetics Systems · 2025
The problem of deployment on hardware platforms, also called mapping problem, is one of the design challenges faced by system architects and engineers during conceptualization. Most of the time, this type of problem is modeled in a flat form by a set of equations and inequalities based on a set of variables. This flat model is then translated into a programming language and solved using a library of solvers. This approach suffers from a number of limitations, particularly when hardware resources become heterogeneous: lack of readability, reuse, and extensibility of the models. By delegating the resolution to an external solver, the link between the problem model and its projection in the solver is lost. The goal of this article is to show how with the recent DEPS problem modeling language (v1) and its integrated modeling and solving environment DEPS Studio it is possible to overcome these limitations. The approach is illustrated by an industrial case study consisting in a deployment of a data-flow application on real-time sensor subsystems, composed of heterogeneous hardware resources.