
Re: 21st Century 'Hilbert' problems
Posted:
Feb 6, 1999 9:05 AM


Simon Croall wrote: > > Hi, > > With the 21st century almost upon us; > > In a similar fashion to Hilbet's 23 problems but forward in 1900. What > would be the problems that you think should be presented to challenge > mathematicians of the 21st century? >
Come on guy's  no one seems to be answering seriously !!!
OK, here is a problem. I am not quite sure that it attains to quite the standards of Hilbert's problems, but I think that it is a fascinating problem. I think that it goes back to Morrey. Basically it asks whether rankone convex functions are quasiconvex. Let me elaborate.
Suppose that U(x,y,z,w) is a function of four real numbers so that the function
t mapsto U(x+ta,y+tb,z+tc,w+td)
is convex whenever the vector (a,b) is parallel to (c,d).
Show that if f(x,y) and g(x,y) are smooth functions with compact support on R^2, and that U(0,0,0,0) = 0, then
int int U(f_x(x,y),f_y(x,y),g_x(x,y),g_y(x,y)) dx dy >= 0
where f_x(x,y) represents partial derivative of f(x,y) with respect to x, etc.
In a sense, this conjecture could be seen as extending Jensen's inequality.
There are counterexamples of higher dimensional analogues due to Sverak. Well, if anyone is really interested, Al Baernstein and myself wrote a paper with very partial results. Look at http://math.missouri.edu/~stephen/preprints/ the paper called:Some conjectures about integral means of $\partial f$ and $\overline{\partial} f$ It also has lots of references to the work of others.

Stephen MontgomerySmith stephen@math.missouri.edu 307 Math Science Building stephen@showme.missouri.edu Department of Mathematics stephen@missouri.edu University of MissouriColumbia Columbia, MO 65211 USA
Phone (573) 882 4540 Fax (573) 882 1869
http://math.missouri.edu/~stephen

