My understanding is that those islands have underground tunnel networks, much like Taiwan proper.
I.e. there are multiple portals(active, hidden), where launchers can pop out and then fire it a matter of seconds, and then run back into the tunnel. As long as there are unsealed portals or as opening work is effective, or as long as there's surviving ammo inside - those can be expected to fire.
For preemptive strike, what i meant to situation similar to Iran: when(if) island gov decides that invasion is imminent and they're past the point of deterrence, they can strike first. For clear numerical reasons, each HIMARS allows x3...x6 more volume of fire per salvo per vehicle.
Taiwan itself is ultimately far, and they don't produce their own oversized rocket artillery. Moving launchers closer is the only way to increase volume with existing means.
That is just a generic "fire the first shot" strategy, which is... fine, but it's also something that the ROC military itself could try to do from Taiwan proper, and they do have some other domestic long range fires to do it with (though of unknown stocks).
Moving some MLRS to Penghu will gain an additional 50-70km of westward reach, but at the expense of putting your platforms on a much smaller piece of real estate (than if they were held on Taiwan island), which will be subject to the same totality of PLA ISR and fires as Taiwan proper -- which recall, the PLA ISR and fires system of systems in a Taiwan conflict is procured and developed around striking much more systems on Taiwan itself, with more surface area and more hiding places, at longer distances.
If the goal is to gain an extra 50-70km of westward reach for HIMARS in particular, that is fine, but in context of the overall system of systems air-naval-missile conflict and associated ISR and EW involved, the much smaller surface area of Penghu (versus Taiwan island) will likely also require a far smaller amount of PLA resources to neutralize the launchers and in a faster manner, than if they were held on Taiwan island itself to survive until more permissive targets ventured closer to Taiwan island (such as to remain hidden as best they can until the PLA conduct an amphibious assault).
But if the goal is to be a sacrificial glass cannon with the aim to try and pull a trigger faster than the other guy, I suppose that's also a strategy, though questionable on the large scale of the cross strait military balance.