Never spend a long time creating a model and then expect to get results out at the last minute (i.e. don't spend all week modelling and expect to get the correct results on Friday afternoon). Instead, iterate (design is iterative), do a bit of modelling then try a test simulation or two before refining the model. (This can also be true for most other areas of simulation). These test simulations should be inexpensive (i.e. Low/Medium quality). Don’t waste time waiting for a complex simulation, which you are going to throw away once you have refined the model. If you want a more complex simulation use what would be non-productive time to run the simulation i.e. lunchtime or overnight. Test simulations should be tidied up (i.e. deleted) when you have an improved simulation.
