Once you have created a few model elements, entered a few key data and created some Resource Requirements, you will quickly reach a point where your model will run, even if the majority of the detailed model development still lies ahead. It is generally instructive to run a model as soon as possible, to verify that the basic structure and results are correct, before the calculations are complicated by additional, advanced inputs. On the same basis, it continues to be prudent to run a model from time to time as it is developed, so that if the results suddenly change dramatically, it is not too difficult to work out the cause.
- Select Save and Run from the File menu. The model data and views are saved. (Alternatively, pressing <F5> will save and run the model). If the model has not been saved before, the Save Model dialog will be displayed, so that you can supply a name for the model – see 4.2.1 Saving a model for the first time.
- The model data is then checked for any errors or inconsistencies. If there are any errors, the source of the first of these is then displayed in the Editor, as described in 4.22 Finding errors in a STEM model.
- If there are just some warnings, e.g., a Service has no Resource Requirements, then you can choose to examine these in the Editor, as if they were errors, or run the model regardless.
- Otherwise STEM will proceed to run the model. The full cycle of STEM algorithms is performed for each year of the model run, in order to produce the results. These are then sorted, prior to being loaded into the Results program, where you can browse the results and check the integrity of your model.
Note: If the model has not been modified since it was last saved, the Save and Run command just appears as Run on the File menu.
Running scenarios and sensitivities
When you add senarios and sensitivities to a model, controlling what exactly you want to run is controlled in the Scenarios and Sensitivities dialog – see 9.3.5 Saving and running scenarios.
Finding run-time errors
Although STEM performs a thorough check of the input data before running a model, as described in 4.22 Finding errors in a STEM model, there are some more complicated problems which can only be detected when the model is actually running, e.g., the coincidence of incremental demand, and rip-out for a resource.
Under these circumstances, the STEM algorithms cannot be continued satisfactorily and the model run will be abandoned, with a message explaining the nature of the problem, which you will then have to remedy in the Editor before running the model again.