Yes, that sounds like a good idea: you could subtract them,
and then determine in which intervals the function h(x) is greater than zero and where it is less than zero. For that you need to determine the zeros of h:

such that
You could do this numerically, for example, using a fixed point iteration method such as Newton-Raphson:
initial guess:
REPEAT
UNTIL

,
where Eps is a tolerance, Eps = 10^{-4} (for example)
In general choosing an adequate initial value sufficiently close to the zero, x
0, so that the method converges is not an easy task, and neither is determining how many zeros a nonlinear function has :S