Vol. 15, No. 1, 1965

Download this article
Download this article For screen
For printing
Recent Issues
Vol. 294: 1
Vol. 293: 1  2
Vol. 292: 1  2
Vol. 291: 1  2
Vol. 290: 1  2
Vol. 289: 1  2
Vol. 288: 1  2
Vol. 287: 1  2
Online Archive
The Journal
Editorial Board
Special Issues
Submission Guidelines
Submission Form
Author Index
To Appear
ISSN: 0030-8730
Concerning periodic subadditive functions

Robert Franklin Jolly

Vol. 15 (1965), No. 1, 159–169

The author investigates those subsets M of the complex plane with the group property that M is closed with respect to complex multiplication. In particular if M is closed, bounded and has for its boundary a curve given in polar form by ρ(θ) = r(θ)exp(iθ) where r is a positive continuous function with period 2π, then r is characterized by these requirements, together with the additional condition that r be submultiplicative. If f(x) = logr(x), the corresponding conditions on f are: f is a continuous nonnegative subadditive function with period 2π.

Some relations between the roots (zeros) and periods of subadditive functions are discussed and in particular, it is shown that: if f is a continuous subadditive function not identically zero, with period 1 and with a root c (i.e., f(c) = 0), then c is a rational number mn (in lowest terms), f(0) = 0 and f has period 1n.

For each positive number c and function f on the set of all numbers, a type of polygonal approximation P(c,f) is defined such that if f is continuous, limP(c,f) = f uniformly over every bounded number set as c 0. If f is subadditive, P(c,f) is subadditive. The subadditive P(c,f) are characterized in terms of their slopes. Since a change of scale does not affect the subadditive property, the author studies functions with period 1 rather than those with period 2π. For each positive integer n, the collection Kn of all functions P(1n,f) for all continuous subadditive functions f with period 1, is shown to have a finite basis. In fact, Kn forms a function cone with finitely many extremal elements (the basis). While an explicit representation is not given, the proof shows how these extremal elements may be constructed.

Several examples are given to illustrate some pathological cases. The methods of this paper may easily be applied to the solution of certain other functional inequalities with corresponding restrictions.

Mathematical Subject Classification
Primary: 26.00
Received: 5 March 1964
Published: 1 March 1965
Robert Franklin Jolly