While the total length of the average Microwave’s wave is about 4.7 - 4.9 inches (12.5 CM) you can further pinpoint the phase of the wave as well both by frequency (playing with that .2 inches in the bandwidth) and phase modulation. This could be further tuned if needed by allowing Microwave ovens to operate in the other ISM band of 5.7 GHZ allowing for 2 inch waves (5.3 CM) or even the 61.25 GHZ band (0.19 inches). Though, as you move up in frequency, you see less penetration as the power is lost faster on the surface of the objects.
Would any of that really make it heat more efficiently though? You’d need at least two magnetrons, some sort of computer vision system, and a computer to do the necessary calculations. Even if you could practically produce an interference pattern that’s better than a single standing wave, I suspect you’d lose more energy than you save.
