Simulate [source] is to model a complex system to predict its behavior in the real world.
This is an approach to represent the operation of a real system consists of several entities, to model the various interactions between them, and finally evaluate the overall behavior of the system and its evolution over time.
The use of simulation overcomes the limitations of the complexity of analytical models.
However, it is necessary to identify the characteristics of the system to represent, as accurately as possible, by abstract models.
If the representation of the real system by abstract models is sufficiently accurate and realistic, it is possible to postpone the results obtained with these models to the real system.

The simulation of a real system [source] becomes necessary when analytical models become either too complex in terms of computation time and resolution, is too simplified vis-�-vis the reality of making, by this fact, the results unrepresentative of the system behavior in a real environment.
Thus, the simulation may be necessary in the following cases:
-The system is not decomposable into subsystems simple and independent of one another, making a very complex analytical modeling.
-The system does not yet exist.
-Experiments on real system is too costly in terms of material and human resources.
-Experiments on real system are not reproducible or representative of all possible environments.