Tom K. answered 06/08/20
Knowledgeable and Friendly Math and Statistics Tutor
These theorems are well-known results;
for the first, we consider g(x) = f(x) - x
Since f(x) is in [a,b] on [a,b], f(a) >= a and f(b) <= b.
If either of these is an equality, then that is a fixed point, so you assume otherwise.
If f(a) > a, then g(a) = f(a) - a > 0
If f(b) < b then g(b) = f(b) - b < 0
If f is continuous on [a,b], so is g
As g(a) > 0, g(b) < 0, and g is continuous, g must take on all values between g(a) and g(b).
This includes 0. When g(x) = 0, f(x) = x, so we have a fixed point.
The uniqueness of the fixed point if the derivative exists and is less than one on the interval can be proven using the Mean-Value Theorem (in fact, the contrapositive).
Let there be a fixed point at y and z in [a,b]
Then, f(y) = y and f(z) = z
Then, the Mean-Value Theorem tell us that there must exist a c in [y,z] such that f'(c) = (f(z) - f(y))/(z-y) = (z-y)/(z-y) = 1.
Thus, there must be a point c in [y,z] where the derivative = 1. Yet, we are given that the derivative is less than one everywhere, which is a contradiction. Thus, there must be at most one fixed point (which we have proven exists), so there is a unique fixed point.
Then, the result that there is a unique fixed point on [0, 2 π] for f(x) =π + 1/2 sin(x/2) can be easily shown.
As sin ranges from -1 to 1, π + 1/2 sin(x/2) ranges from π - 1/2 to π + 1/2. As 0 < π - 1/2 and π + 1/2 < 2 π,
f(x) is bounded by [0, 2 π] on [0, 2 π], so there exists a fixed point.
f'(x) = 1/4 cos(x/2), which is <= 1/4 < 1 everywhere
Thus, the fixed point is unique.

Tom K.
where g(x) = 0, f(x) has a fixed point. As f(x) is between a and b throughout the interval, g(x) is >= 0 at the left endpoint and is <= 0 at the right endpoint; if we have exactequality at the endpoint, we are done. If not, g is greater than 0 at the left endpoint and less than 0 at the right endpoint. As g is continuous, if it goes from being above 0 to being below 0, at some point, it has to equal 0.06/10/20
Jamie B.
Tom if you have a moment--how would you explain that the ad hoc utilization of this convenient function [g] = [f] -x leads to a proof for all functions [g] so described in the question (cheers, btw: this is the second competent response Tom has beat me to by a matter of minutes--ya rat !)06/08/20