Page 56 -
P. 56

V. The Waring Problem
                        50
                        A. Theorem (Dirichlet). Given a real x and a positive integer M,
                        there exists an integer a and a positive number b ≤ M such that
                              a       1
                        |x −   |≤         .
                              b     (M+1)b
                        Proof. Consider the numbers 0, x,2x,3x, ..., Mx all reduced
                        (mod 1). Clearly, two of these must be within   1   of each other.
                                                                       M+1
                        If these two differ by bx, then 1 ≤ b ≤ M and bx (mod 1) is, in
                                        1
                        magnitude, ≤       . Next pick an integer a that makes bx − a equal
                                       M+1
                                                       1                    a        1
                        to bx (mod 1). So |bx −a|≤        which means |x − |≤           ,
                                                     (M+1)                  b     (M+1)b
                        as asserted.                                              Q.E.D.
                           We also point out that this is a best possible result as the choice
                               1
                        x         shows for every M. (Again, we may assume that (a, b)
                              M+1
                        1 for, if they have a common divisior, this would make the inequality
                        |b|≤ M even truer).


                        B. Schnirelmann’s Theorem. If S is a set of integers with positive
                        Schnirelmann density and 0 ∈ S, then every non-negative integer is
                        the sum of at most k members of S for some k ≥ 1.


                        Lemma 1. Let S have density α and 0 ∈ S. Then S ⊕ S has density
                                       2
                        at least 2α − α .
                        Proof. AllthegapsinthesetS arecoveredinpartbythetranslation
                        of S by the term of S just before this gap. Hence, at least the fraction
                        α of this gap gets covered. So from this covering we have density α
                        from S itself and α times the gaps. Altogether, then, we indeed have
                                                 2
                        α + αà 1 − αð   2α − α , as claimed.

                        Lemma 2. If S has density α>       1  , then S ⊕ S contains all the
                                                           2
                        positive integers.

                        Proof. Fix an integer n which is arbitrary, let A be the subset of
                        S which lies ≤ n, and let B be the set of all n minus elements of
                        S. Since A contains more than n/2 elements and B contains at least
                        n/2 elements, the Pigeonhole principle guarantees that they overlap.
                        So suppose they overlap at k. Since k ∈ A,weget k ∈ S, and since
   51   52   53   54   55   56   57   58   59   60   61