It's for formal verification and validation (V&V).
> why the development of a simulation should hold up the full production
Why shouldn't it hold up production? The program has requirements for formal V&V that they haven't met yet. In theory, you could start production by accepting the risk that you might need to do some rework, but I don't know if that was even an option.
> step through iterations with very quickly
The argument for doing a simulation is to get higher fidelity. It doesn't really make the process any faster on its own. The simulations themselves take a trivial amount of time compared to the preparation and analysis.
> nothing like this has been done before
There's a lot of simulation in aerospace and defense. I don't know exactly how JSE compares in terms of extent and reliance on it, but it's not a completely new idea.
JSE is really interesting, it's meant to be a high-fidelity electronic warfare simulation with actual aircraft software in the simulation loop. The F-35 that's present in the JSE is the same code that is running on a real jet, with sensor models specifically developed by all the sensor vendors. All this is fed with the latest intelligence data on threat aircraft and emitters. The point of all this is, the US should be able to better develop the aircraft -- and other aircraft types -- in the future if it has a high tech digital testing ground that it can step through iterations with very quickly.
As far as I can tell nothing like this has been done before so it's no wonder that it's taking a while.
Here's a US Navy PDF briefing on the JSE: https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2018/te...