The function #f(x) = 1/(1-x)# on #RR# \ #{0, 1}# has the (rather nice) property that #f(f(f(x))) = x#. Is there a simple example of a function #g(x)# such that #g(g(g(g(x)))) = x# but #g(g(x)) != x#?

1 Answer
Jul 10, 2015

The function:

#g(x) = 1/x# when #x in (0, 1) uu (-oo, -1)#

#g(x) = -x# when #x in (-1, 0) uu (1, oo)#

works, but is not as simple as #f(x) = 1/(1-x)#

Explanation:

We can split #RR# \ #{ -1, 0, 1 }# into four open intervals #(-oo, -1)#, #(-1, 0)#, #(0, 1)# and #(1, oo)# and define #g(x)# to map between the intervals cyclically.

This is a solution, but are there any simpler ones?