...
"A total of about 1366 watts/sq.meter of solar radiation (all wavelengths) reaches earth from the sun. "

I don't think 1366 watts for all wavelength is correct (my ref says 1418 watts) but for sunlight energy which can be converted, I would accept the 1366 watts number.

"Only a fraction (~0.18) penetrates the atmosphere to ground level in the midwest USA."
This is WAY off. Just off hand I would say that on a clear day and in the afternoon more than 90% of the solar energy (visable light) penetrates the atmosphere to ground level.

"This means that a sunlight collector orbiting earth only has to be ~0.18 times the size of the greenhouse acreage. "

The major factor which would increase total solar energy would be the increase time of exposure to solar energy (no or little night and no clouds/morning haze, etc). In the best location in orbit you could get twice the solar energy as best place on the Earth surface, but not 5 times as much.
Unless you are referring to ultraviolet light- which a large portion is blocked by Earth's ozone. I suppose it might be possible to genetically engineer plants so that they could survive (thrive) under higher levels of UV.
Hmm, that's pretty interesting come to think about it. I know plants need UV to grow and I know too much UV will kill plants. But I don't know what type of light plants need to grow- in other words if you just give plants only UV light (and the right amount) would plants grow without the visible portion of the light spectrum. That's a pretty simple question, I wonder if anyone knows the answer.
?