How many parallel approaches to an engineering problem/model are you typically working on?

If you have a project you're working on and you're trying out different approaches for the solution, how many parallel models for the same thing are you typically working on?
Have a couple of files to work on and see which one works better in the end?
Try to create one version from the beginning to the end before starting to work on something alternative?
Or do you think about the alternative solutions only in the head and pick the best approach already before modelling?

4 Answers

Usually i adapt sketches and mental thoughts to a scrapbook first.

Choose my direction and follow it through with maybe two variations.

Or if i feel the need to test something a different way just save as the model to test my desired configuration. If both approaches work i usually create configurations of my model and then run analysis after completion such as fea/prototyping etc...

Only one model...
...changes goes during modeling itself...When you work with two or more same projects at the same time (as I did earlier), you risk to getting lost in mass of problems and nervous due lack of solutions...I realize this (used "save as" button a lot) when none of "save as" projects never open again and work is done...
With one project u are always in hot spot with any problems involved...
...this is my way, try it and than build your own...
Good luck :)))

Depends largely on the amount of time available.
Much time, many approaches with different methods.
No time, i spend some time thinking about the best method an then go for that.

My CAD models evolve during design. I usually work on one, but keep 3 generations active so I can refer back. At the end I only keep one model.