The most likely solution, and certainly the one I would use is deck load the Enterprise and Hornet with the Yorktown's air wing and then fly them off to Midway to operate from land. I would then alter the US plan to allow a closer Japanese approach before a counterstrike as the Japanese don't know the US carriers are present.

The likely result is still three Japanese carriers sunk. But, even if the US just cripples the Kido Butai through massive losses of aircraft and crews the result is still largely the same too. Given Midway had radar and the Navy fighter aircraft would be just as effective flown from land the Japanese could expect as much as 50 to 75% losses among their strike aircraft.

The US strikes on the Japanese carriers worked mainly becasue the Japanese CAP doctrine and coordination was so poor. Lacking radar they relied almost entirely on their fighter CAP spotting and intercepting raids. This typically happened at less than 20 miles from their carriers giving them little time to work a raid over.
With a large CAP they were able to put crippling losses on the US torpedo planes that showed up first. But, as a result of this their CAP was not set to take on a second raid at high altitude and there was no means of detection (eg., radar) and poor communications to call the fighters back in a timely fashion. The result was the loss of three carriers.
Compounding this was the lack of a coherent air defense doctrine by their surface ships. If you look at the Japanese deployment it was one of maneuver by column of divisions much as WW 1 battleships would do.

Anyway, the Yorktown's air group could have launched a strike from Midway just as easily as from the flight deck of a carrier. Deck loading them on the other carriers and striking down their own air wings was not only possible but easily doable as US doctrine did not normally call for hangering the air wing so there was plenty of space to do so while deck parks were the norm.