That's an interesting question. Power required for an airplane flying level and not accelerating goes up with the cube of airspeed. In this case, though, I take it that you are holding airspeed constant and allowing thrust to vary. Sounds like you are figuring a delta RPM command for a power-compensation circuit, you rascal. I would guess that for constant blade angle of attack, thrust losses would be proportional to the square of RPM, but angle of attack is sorta inversely proportional to RPM, so torque would be approximately proportional to RPM, hence power to the prop would be approximately proportional to square of RPM. Igor or somebody could give you the real lowdown.
This reminds me of the time I was driving a truck across Wyoming on I-80 with a strong crosswind. I had to hold in a significant steering wheel torque to keep the truck going straight, meanwhile trying to calculate whether the force would get greater if I sped up or slowed down.