Reactor recipe change behavior

[opotowsky]

In probing some unexpected behavior with using the recipe change functionality in the cycamore reactor, I have found the following:

• This line needs to be updated to if (t < change_t) { so that newly deployed reactors "see" the recipe change.
• This functionality only works successfully when a singular source with no specific recipe is used to supply the commodity that undergoes a recipe change. If two recipe-specific sources are used to supply the same commodity, the second facility to bid will always win no matter which recipe it has.
• Perhaps the above bullet point is OK, but this limitation needs to be noted in the documentation if so.
• This also causes an interesting problem. If the reactor sends off spent fuel in a timestep after the recipe change but before the refuel, the NEW spent fuel recipe is used even though it does not follow from the new input fuel recipe

PR #533 addresses the first bullet point for the preference change functionality in the reactor, but since the recipe change functionality needs more work the update is not included in that PR.