The field is principally devoted to consideration of
direct problems over (typically) the integers, that is, determining the structure of from the structure of : for example, determining which elements can be represented as a sum from , where '''' is a fixed subset. Two classical problems of this type are the
Goldbach conjecture (which is the conjecture that contains all even numbers greater than two, where is the set of
primes) and
Waring's problem (which asks how large must '''' be to guarantee that contains all positive integers, where :A_k=\{0^k,1^k,2^k,3^k,\ldots\} is the set of th powers). Many of these problems are studied using the tools from the
Hardy-Littlewood circle method and from
sieve methods. For example, Vinogradov proved that every sufficiently large odd number is the sum of three primes, and so every sufficiently large even integer is the sum of four primes.
Hilbert proved that, for every integer , every non-negative integer is the sum of a bounded number of th powers. In general, a set ''
of nonnegative integers is called a basis
of order if contains all positive integers, and it is called an asymptotic basis
if contains all sufficiently large integers. Much current research in this area concerns properties of general asymptotic bases of finite order. For example, a set is called a minimal asymptotic basis'' of order '
if ' is an asymptotic basis of order but no proper subset of '
is an asymptotic basis of order '. It has been proved that minimal asymptotic bases of order '
exist for all ', and that there also exist asymptotic bases of order '
that contain no minimal asymptotic bases of order '. Another question to be considered is how small can the number of representations of as a sum of '''' elements in an asymptotic basis can be. This is the content of the
Erdős–Turán conjecture on additive bases. ==See also==