The inhomogeneous minimum of binary quadratic forms

An indefinite binary quadratic form ƒ gives rise to a certain function M on the torus. The properties of M, especially those related to its maximum – the so-called inhomogeneous minimum of ƒ – are the subject of numerous papers. Here we continue this study, putting more emphasis on the general behaviour of M.

A Restricted Inhomogeneous Minimum for Forms

Let us suppose that ƒ(x, y) is an indefinite binary quadratic form that does not represent zero. If P is the real point (x0, y0) then we may define where the infimum is taken over all integral x, y. The inhomogeneous minimum of the form ƒ is defined where the supremum taken over all real points P, need only extend over some complete set of points, incongruent mod 1.

An Inhomogeneous Minimum For Non-Convex Star-Regions With Hexagonal Symmetry

1. Introduction. Several authors have proved theorems of the following type:Let x0, y0 be any real numbers. Then for certain functions f(x, y), there exist numbers x, y such that1.1 x ≡ x0, y ≡ y0 (mod 1),and1.2 .The first result of this type, but with replaced by min , was given by Barnes (3) for the case when the function is an indefinite binary quadratic form. A generalisation of this was proved by elementary geometry by K. Rogers (6).

On the quadratic form x 2 —7 y 2

1. Let ƒ( x,y ) = ax 2 + bxy + cy 2 be an indefinite binary quadratic form, with real coefficients a, b, c , and discriminant d = b 2 —4 ac . A well-known theorem of Minkowski asserts that for any real numbers x 0 , y 0 there are integers x, y such that f(x + x0,y + yQ) (!) ♦Various authors (Heinhold 1938, 1939; Davenport 1946; Varnavides 1948 a ) have given results which are more precise than this. In particular, their results give, for a few special quadratic forms, the exact value of the constant, i.e. the least number by which ¼ √ d may be replaced on the right of (1). Perhaps the simplest form for which the existing results do not determine the exact value of the constant is the form x 2 — 7 y 2 with discriminant 28. Here ¼ √ d = ½ √7 and this was improved by Heinhold (1938, p. 683) to ¾. In this paper, we attack the problem by a method based on that of Davenport (1947), and find the exact value of the constant. We prove that it is always possible to satisfy

The minima of a pair of indefinite, harmonic, binary quadratic forms

If f is a real, indefinite, binary quadratic form of discriminant d and if κ(f) is the minimum of | f | taken over all integer values of x, y, not both zero, then it is well known that and that this is a ‘best possible’ result.