Page 98 - DMTH401_REAL ANALYSIS
P. 98

Real Analysis




                    Notes          As an application of the above result, we will see that perfect sets are closed sets that contain lots
                                   of points:

                                   8.3.1 Perfect Sets are Uncountable

                                   Every non-empty perfect set must be uncountable.

                                   Proof: If S is perfect, it consists of accumulation points, and therefore can not be finite. Therefore
                                   it is either countable or uncountable. Suppose S was countable and could be written as
                                                                  S = {x , x , x , ...}
                                                                       1  2  3
                                   The interval U  = (x  – 1, x  + 1) is a neighbourhood of x . Since x  must be an accumulation point
                                              1   1    1                      1      1
                                   of S, there are infinitely many elements of S contained in U .
                                                                                   1
                                   Take one of those elements, say x  and take a neighbourhood U  of x  such that closure (U ) is
                                                              2                       2    2                2
                                   contained in U  and x  is not contained in closure (U ). Again, x  is an accumulation point of S, so
                                              1     1                       2       2
                                   that the neighbourhood U  contains infinitely many elements of S.
                                                        2
                                   Select an element, say x , and take a neighbourhood U  of x  such that closure (U ) is contained in
                                                     3                       3   3                3
                                   U  but x  and x  are not contained in closure (U ).
                                    2    1     2                        3
                                   Continue in that fashion: we can find sets U  and points x  such that:
                                                                      n          n
                                      closure (  ) 
                                                n + 1  n
                                      x is not contained in   for all 0 < j < n
                                        j                 n
                                      x  is contained in 
                                        n              n
                                   Now consider the set
                                      V = (closure ( )  S)
                                                     n
                                   Then each set closure ( )S) is closed and bounded, hence compact. Also, by construction,
                                                       n
                                   (closure (  )S) (closure ( )S). Therefore, by the above result, V is not empty. But which
                                           n + 1           n
                                   element of S should be contained in V? It can not be x , because x  is not contained in closure (U ).
                                                                            1        1                       2
                                   It can not be x  because x  is not in closure ( ), and so forth.
                                              2        2               3
                                   Hence, none of the elements {x , x , x , ...} can be contained in V. But V is non-empty, so that it
                                                            1  2  3
                                   must contain an element not in this list. That means, however, that S is not countable.
                                   8.4 Cantor Middle Third Set

                                   Start with the unit interval

                                   S  = [0, 1]
                                    0
                                   Remove from that set the middle third and set
                                   S  = S \(1/3, 2/3)
                                    1  0
                                   Remove from that set the two middle thirds and set

                                   S  = S \{(1/9, 2/9) (7/9, 8/9) }
                                    2  1
                                   Continue in this fashion, where
                                   S  = S \{middle thirds of subintervals of S }
                                    n+1  n                            n
                                   Then the Cantor set C is defined as

                                   C = S
                                         n


          92                                LOVELY PROFESSIONAL UNIVERSITY
   93   94   95   96   97   98   99   100   101   102   103