arXiv++ Combinatorics

Consider a set of integers $\mathscr A$ having finite diameter $X$, and a system of simultaneous polynomial equations to be solved over $\mathscr A$. In many circumstances, it is known that the number of solutions of this system is $O(X^ε|\mathscr A|^θ)$ for a suitable $θ>0$ and any $ε>0$. These estimates become worse than trivial when the diameter $X$ is very large compared to $|\mathscr A|$, or equivalently, when the set $\mathscr A$ is very sparse. This motivates the problem of seeking a new set of integers $\mathscr B$, in a certain sense isomorphic to $\mathscr A$, having the property that the diameter $X'$ of $\mathscr B$ is smaller than $X$, and at the same time the set $\mathscr B$ preserves the salient features of the solution set of the system of equations in question. We report on our speculative investigations concerning this problem closely associated with the topic of Freiman homomorphisms.

We prove that any finite set of real numbers can be split into two parts, one part being highly non-additive and the other highly non-multiplicative.

We describe a method for bounding the set of exceptional integers not represented by a given additive form in terms of the exceptional set corresponding to a subform. Illustrating our ideas with examples stemming from Waring's problem for cubes, we show, in particular, that the number of positive integers not exceeding N, that fail to have a representation as the sum of six cubes of natural numbers, is O(N^{3/7}).