UNICEF-julkort med universitetets emblem finns hos Birgitta Lumsden.
From pol@math.kth.se Thu Nov 30 14:07 MET 1995 Return-Path: <pol@math.kth.se> Received: from newton.math.kth.se by math.liu.se (5.x/SMI-SVR4) id AA00561; Thu, 30 Nov 1995 14:07:54 +0100 Received: from karush.math.kth.se by newton.math.kth.se (5.65+bind 1.7+ida 1.4.2/4.0) id AA05614; Thu, 30 Nov 95 14:07:55 +0100 Received: by karush.math.kth.se (5.65+bind 1.7+ida 1.4.2/6.0) id AA00205; Thu, 30 Nov 95 14:06:53 +0100 From: P O Lindberg <pol@math.kth.se> Date: Thu, 30 Nov 95 14:06:53 +0100 Message-Id: <9511301306.AA00205@karush.math.kth.se> To: malin@math.liu.se Subject: nytt forsokmigen Content-Type: text Content-Length: 1113 Status: R
Abstract for Colloqium 15 dec.
P O Lindberg
On functional dependence
In most classical analysis books there are results on functionla
dependence. Two functions f and g are functionally dependent if there
is a nontrivial function F such that 
is identically 0.
In couple of practical applications, we have come across situations
where we would like to conclude that two functions are functionally
dependent globally in the more narrow sense that 
for some
function h. Then the classical results are not so useful, because they
are local. Moreover they are usually stated in terms of functional
determinants rather than in terms of notions with more geometric
content.
Our simple functional dependence result reads (essentially), that if
there is a function 
such that
and g has connected level surfaces and del 
has full rank, then
for some h.
We will sketch the proof of this result and indicate how it was used in one or two cases:
homothetic functions and invariance of achieved utility in additive random utility models.