Complexity of Simulation Modeling

Although modeling has been used with great success, many organizations have been wary of using modeling tools. Part of this reticence is because the early modeling tools were complicated, labor-intensive, took a long time to produce a result, and had questionable accuracy.

Every management team that uses modeling tools should therefore have a clear idea of what constitutes a good enough solution. By its nature, modeling will always be an approximation. Some variables are omitted to keep the model to a reasonable size. Other variables that affect performance may not even be known or included in the model, and the assumptions about operational conditions add another approximation.

It is always possible to put more work and effort into any model. You can refine the assumptions about operational conditions, adding more details and continually trying to close the gap between the model result and the actual reality. It's very easy to approach the point of diminishing returns, with each improvement in the model's accuracy taking increasingly larger increments of effort and time. At some point, the model has to be declared good enough, unless significant changes dictate more work to bring the model up to date.