When simulation is appropriate or not

Simulation and modelling

Modelling is the process of building an abstract representation (physical or mathematical) of a system. Simulation is the imitation of the operation of that system over time by experimenting with the model, usually on a computer, to understand its behaviour or evaluate alternatives.

When simulation is appropriate

• To study and experiment with the internal interactions of a complex system or a subsystem.
• To see the effect of informational, organizational and environmental changes by changing the model and observing results.
• When knowledge gained during model building suggests improvements in the system.
• To find which variables are most important and how they interact.
• To test new designs, layouts or policies before spending money on them.
• To verify analytical solutions.
• To train operators (flight simulators) and to demonstrate a design through animation.
• When time must be compressed or expanded to study a system.
• When the real system is too dangerous, costly or slow to experiment with.

When simulation is not appropriate

• When the problem can be solved by common sense.
• When it can be solved analytically (e.g. a simple M/M/1 queue formula).
• When direct experiment on the real system is easier and cheaper.
• When the cost of simulation exceeds the likely savings.
• When resources or time (people, computers, deadline) are not available.
• When data are not available to build, verify and validate the model.
• When not enough time or skill exists to verify and validate the model.
• When managers have unreasonable expectations, such as expecting too much too soon or exact answers from a simulation.
• When the system behaviour is too complex or cannot be defined, for example human behaviour that is hard to model.