The paper deals with counting lattice points, which means points with whole-number coordinates, inside geometric shapes. Specifically, it studies triangles living in three-dimensional space. An Ehrhart polynomial is a formula that counts how many lattice points fall inside (or on the boundary of) a shape when you scale it up by a whole-number factor. The authors develop two independent methods for finding these counting formulas for triangles in 3D space, drawing on classical results from number theory and geometry. One approach uses a theorem about remainders and division (Popoviciu's Theorem), and the other adapts a well-known area formula for lattice polygons (Pick's Theorem) to work in a 3D setting where the triangle does not lie flat in a coordinate plane.
The authors then connect this geometric machinery to a practical problem in number theory: counting the number of ways to solve equations of the form ax + by + cz = n, where a, b, c, and n are fixed positive integers and you want solutions where x, y, and z are all non-negative whole numbers. Each such solution corresponds to a lattice point on a triangle in 3D space, so the Ehrhart polynomial approach directly gives a counting tool. Using their two methods, they derive both upper and lower bounds on the number of solutions, and these bounds are given by explicit quasi-polynomials, which are formulas that cycle through a repeating pattern of expressions depending on the remainder when n is divided by certain numbers. Crucially, the gap between the upper and lower bounds stays bounded by a fixed constant that depends only on a and b, not on n.
As a concrete payoff, the paper proves a result about the Frobenius problem, sometimes called the coin problem. The classical version asks for the largest number that cannot be represented as a non-negative integer combination of given coin denominations. The authors generalize this to the k-th Frobenius number, which is the largest value of n for which the equation ax + by + cz = n has fewer than k non-negative solutions. They prove that for any three positive integers a, b, c and any positive integer k, this k-th Frobenius number always exists, meaning there is always a largest n beyond which every value can be represented in at least k distinct ways.