Recall that inverse functions only exist for invertible (one-to-one) functions. Otherwise, the functions needs to be split into domains of invertible functions, called branches. The desired domain may be specified in the question or by the context.
Also if question says an inverse exists, then it exists without having to break up into branches. This allows you to find domain and range of the inverse without having to find the inverse itself.
For inverse of quadratics, a simple ± is sufficient to describe the two branches.
Example: Given f(x)=9x2+12x−3,x≤−1. Find f−1(x) and state the domain.
arcsin and arctan are increasing functions, from an increasing interval of sin and tan, while arccos is a decreasing function, from a decreasing interval of cos.
If sin, cos, or tan is vertically reflected across x-axis, intervals of increasing function becomes intervals of decreasing functions, vice versa. cos is unaffected by a horizontal reflection across y-axis, but the other two are, with similar increasing to decreasing changes.
So if looking for the inverse function of a decreasing portion of sin or increasing portion of cos, then we need to multiply the inverse by −1.
Conceptually, we identify transformations, then convert vertical stretches and translations to horizontal stretches and translations in the inverse, and vice versa.
Example: Given
f(x)=−3cos(2(x+3π))+1,x∈[67π,35π]
Find f−1(x).
See method at transformation of trig functions, there is a maximum at (67π,4) and a minimum at (35π,−2). This confirms f is decreasing.
A simplistic approach for the inverse function yields
x↦2arccos(−3x−1)−3π
While f is decreasing, it also had a vertical reflection, meaning it was transformed from an increasing branch in x↦cosx. As arccos is for a decreasing branch of cos, we need to multiply the inverse by −1. We can ignore the vertical shift −3π for now.
g(x)=−2arccos(−3x−1)
Time to figure out that vertical shift in the final answer.
g(4)=−2arccos(−34−1)=−2π
67π−(−2π)=35π
This means we need to add 35π to −2π to get 67π.