14.02.2013 Views

Gruber P. Convex and Discrete Geometry

Gruber P. Convex and Discrete Geometry

Gruber P. Convex and Discrete Geometry

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Convex</strong> Functions<br />

<strong>Convex</strong> functions came to the face rather late. Early contributions are due to Stolz<br />

[972], Brunn [174], Hadamard [459] <strong>and</strong> Jensen [544] around the beginning of the<br />

twentieth century. The systematic study of convex functions dealing with continuity<br />

<strong>and</strong> differentiability properties, with variants of the fundamental notions <strong>and</strong> with<br />

inequalities began only in the twentieth century. The unpublished seminal lecture<br />

notes of Fenchel [334] in the early 1950s led to convex analysis, a careful study of<br />

analytic properties of convex functions related to optimization. Major contributions<br />

were made by Moreau, Rockafellar <strong>and</strong> Phelps.<br />

<strong>Convex</strong> functions appear <strong>and</strong> are useful in many areas of mathematics, including<br />

the calculus of variations, control theory, inequalities <strong>and</strong> functional equations, optimization<br />

<strong>and</strong> econometrics. On the other h<strong>and</strong>, investigation of convex functions per<br />

se has led to a rich <strong>and</strong> voluminous theory.<br />

This chapter contains basic geometric <strong>and</strong> analytic properties of convex functions<br />

of one <strong>and</strong> of several variables, related to convex geometry. A highlight<br />

is Alex<strong>and</strong>rov’s theorem on second-order differentiability almost everywhere of a<br />

convex function. A further result is of Stone–Weierstrass type. Applications treat<br />

inequalities, the characterization of the gamma function by Bohr <strong>and</strong> Mollerup, <strong>and</strong><br />

a sufficient condition in the calculus of variations due to Courant <strong>and</strong> Hilbert.<br />

For more detailed expositions <strong>and</strong> references on convex functions with an emphasis<br />

on convexity, consult the articles <strong>and</strong> books of Beckenbach [87], Roberts<br />

<strong>and</strong> Varberg [841], Roberts [840] <strong>and</strong> also Giles [379], Van Tiel [1006] <strong>and</strong><br />

Czerwik [234].<br />

We will not treat convex functions in the sense of convex analysis <strong>and</strong> general<br />

analysis. For material in this direction, see Moreau [755], Rockafellar [843] <strong>and</strong><br />

Phelps [800], Giles [379], Van Tiel [1006], Borwein <strong>and</strong> Lewis [158], Hiriart-Urruty<br />

<strong>and</strong> Lemaréchal [505], Magaril-Il’yaev <strong>and</strong> Tikhomirov [678] <strong>and</strong> Niculescu <strong>and</strong><br />

Persson [770].<br />

Throughout the book, vectors are columns, unless stated otherwise, but for convenience,<br />

we write vectors as rows as usual. If there is need to stress, that a row<br />

actually means a column, the superscript T is added.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!