If you want to adjust each map individually instead of against each other like in the other answer you need to default to straight normal that is (0.5, 0.5, 1). That means with values you need to find difference to 0.5 and multiply it with your weight before adding it back to the 0.5 for each R and G channel.
(the weight field of normal map is placed under the respectful image node)
Another option is to mix the colors, again mixing both with the straight normal color (0.5, 0.5, 1) individually with adjustable factors.

Both methods should do the trick but in the color example the input values' magnitude is halved because of the way the colors are combined.