When you address a problem using the STEM modelling system, you are constructing a mathematical model. Mathematical modelling is a way of attempting to represent relationships in the real world as a series of mathematical equations.
The reasons we wish to build a model can be summarised as:
- to provide a structure within which an issue can be explored and understood
- to consider the relationship between quantities
- to provide quantitative evidence to support conclusions
- to provide the assumptions, data and algorithms behind results and conclusions, allowing their validation by others.
A model cannot predict the future, but it can be used to explore scenarios and compare alternatives; it can provide better understanding of the possible consequences, risks and opportunities of strategic options.
With a model you are able to evaluate the robustness of particular strategiesby comparing their performance under a variety of circumstances.
You must therefore be able to check that the model you have built is operating as it was designed and be able to carry out sensitivity analyses to establish how robust your scenario is, and whether small changes in the input assumptions will invalidate your results. If you are not able to estimate one of your input parameters very accurately but you find that it has only a small impact on the output, then you still have a useful model. If, on the other hand, you are not very sure of one of your assumptions and a small change in that assumption makes a big change in the results, then you must either refine the estimating techniques in that area or discard the model.
A model is not the same thing as a modelling system. A model is the system of equations which could theoretically be used to perform the calculations without the use of a computer at all (if there were enough people working on it). The modelling system is the computer program which performs the calculations, and provides an interface to simplify the development of a model.