-% As detailed in Sect.~\ref{sec:mdn}, the mixed case asynchronously combines
-% subsets of synchronized components (the different classes). The double interest
-% of that approach is to ensure the convergence of the system while using
-% asynchronism.
-
-% The part of asynchronism often reduces the global execution time as the
-% communications between subgroups are implicitly overlapped by computations.
-% However, the iterative scheme is no more the same as the synchronous one and its
-% number of iterations to reach the convergence will be greater or equal.
-
-% Le nombre d'itérations requises pour obtenir la convergence en mode mixe
-% dépend des arangements entre les délais de communication et les durées de
-% calcul.
-
-% number directly depends on the arrangement of delays during the execution and
-% then on the communication times. But it also depends on the evolution functions
-% which influence the way each part of the system stabilizes itself.
-% In fact, according to its evolution function, a component may reach its fixed
-% point state even with a part of its input data not recently updated. In
-% addition, as mentioned earlier, the set of components in any system does not
-% stabilize at the same time and there is often a propagation of the stabilization
-% through the system.
-% Also, the previously mentioned phenomenon of stabilization propagation through
-% the system is still present in mixed mode.
