Simulation Explained

What do we mean by Simulation?

Simulation is the science of creating statistically accurate models to represent the behaviour of real life systems in order to subject them to predictive experimentation. These experiments or scenarios can then enable 'what if?' questions to be answered without risk or disturbance to the real life system.

Discrete Event Simulation works by modelling the individual events that occur at given time intervals, taking into account resources, constraints and the way they interact. It also includes the randomness and variability you will see in real life. In this way, it behaves like a real process, such as a manufacturing line; a call centre or an airport baggage handling system. Lanner's WITNESS, L-SIM, PRISM and Px-SIM products all use Discrete Event Simulation.

When would you use Simulation?

Processes that involve variability, disruptions and interaction complexity are ideal for simulation. Other techniques for measuring processes, such as spread sheets and flow diagrams, often ignore or treat these factors as averages, with important consequences.

Let's take an example of an ATM to illustrate the importance of including variability and disruption. It will also illustrate the strength of simulation in providing evidence and in communicating the consequences.

As the graph shows, the consequence of not including variability in this simple system, even without considering disruption such as a machine jamming, is significant. Imagine the consequences when you consider looking to predict the throughput of a new baggage handling system or call centre!

Benefits over real life experimentation

Simulation provides significant advantages over real life experimentation in terms of cost, time and repeatability when variability, disruption and complexity exist. Imagine taking a production line or a hospital accident and emergency unit, then experimenting with various 'what ifs'. It would be impractical and possibly dangerous.

It is however important to point out that organisations that regularly need simulation gain greater benefit when simulation is embedded in a sector specific application, rather than having to skill up to use a general purpose tool. Such applications are pre configured to request the correct data, already incorporate logic to explore the most relevant 'what ifs' and can be used by front line employees. For those organisations that have or are investing in simulation skills, again experience would recommend the selection of a general product that can be extended into an end user application.

Are all simulation software products the same?

There are at least 82 discrete simulation products on the market. Certain packages offer different benefits in different situations, while most packages require a level of scientific background in simulation theory. This situation is changing however. As the simulation market has matured, leading providers have started to deliver products pre configured for particular sectors. These new solutions offer industry specific simulation capabilities, via an intuitive interface that is easy for frontline staff to use.

How to simulate?

Descriptions and videos of simulation products often describe a process of clicking a run button, seeing the clock run and an animation of your process on your screen, then being presented with graphical results. This misses the two most difficult steps; how to represent the real life process within the simulation model and what data is needed to provide meaningful results!

Simulation applications such as Lanner's Px-SIM, PRISM and L-SIM are designed for and focused on a particular sector, so have already done these difficult steps for you. WITNESS whilst able to model virtually any situation with the appropriate skills, is importantly designed to be extended and wrapped into application.

Recommended further reading on best practice to apply simulation

Simulation: The Practice of Model Development and Use: Professor Stewart Robinson, Warwick Business School, University of Warwick

Warwick Business School have a long standing and strong relationship with Lanner, the roots being in the late 1970's and in early research and the application of visual interactive simulation done in conjunction with Lanner.