Page 99 - DMTH401_REAL ANALYSIS
P. 99

Unit 8: Completeness and Compactness




          The Cantor set gives an indication of the complicated structure of closed sets in the real line. It  Notes
          has the following properties:


                 Example: The Cantor set is compact.
          Solution: The definition of the Cantor set is as follows: let
               A   = [0, 1]
                 0
          and define, for each n, the sets A  recursively as
                                    n
                             +
                                   +
                         ¥  æ  1 3k 2 3k ö
               A   = A   /   ç  ,     ÷
                 n   n – 1  k 0è  3 n  3 n  ø
                          =
          Then the Cantor set is given as:
               C = A
                     n
                      +
                            +
                  ¥  æ 1 3k 2 3k ö
          Each set    ç  ,    ÷  is open. Since A  is closed, the sets A  are all closed as well, which
                  k 0è  3 n  3 n  ø           0                n
                  =
          can be shown by induction. Also, each set A  is a subset of A , so that all sets A  are bounded.
                                              n            0             n
                 Example: The Cantor set is perfect and hence uncountable.
          The definition of the Cantor set is as follows: let
               A  = [0, 1]
                0
          and define, for each n, the sets A  recursively as
                                    n
                            +
                                  +
                         ¥  æ 1 3k 2 3k ö
               A  = A  \   ç   ,     ÷
                n   n – 1    n     n
                        k 0è  3   3   ø
                         =
          Then the Cantor set is given as:
               C = A
                     n
          From this representation it is clear that C is closed. Next, we need to show that every point in the
          Cantor set is a limit point.
                                                                                n
          One way to do this is to note that each of the sets A  can be written as a finite union of 2  closed
                                                   n
          intervals, each of which has a length of 1/3 , as follows:
                                             n
               A  = [0, 1]
                0
               A  = [0, 1/3] [2/3, 1]
                1
               A  = [0, 1/9][2/9, 3/9] [6/9, 7/9]  [8/9, 1]
                2
               ...
          Note that all endpoints of every subinterval will be contained in the Cantor set. Now take any
          x C =A . Then x is in A  for all n. If x is in A , then x must be contained in one of the 2  intervals
                                                                               n
                 n            n               n
          that comprise the set A . Define x  to be the left endpoint of that subinterval (if x is equal to that
                             n       n
          endpoint, then let x  be equal to the right endpoint of that subinterval). Since each subinterval
                          n
          has length 1/3 , we have:
                      n
                                          |x – x | < 1/3 n
                                               n
          Hence, the sequence {x } converges to x, and since all endpoints of the subintervals are contained
                            n
          in the Cantor set, we have found a sequence of numbers contained in C that converges to x.

                                           LOVELY PROFESSIONAL UNIVERSITY                                   93
   94   95   96   97   98   99   100   101   102   103   104