Monday, May 10

Quasi-Newton Methods from Davidon to Automatic Differentiation

8:30 AM-9:15 AM
Room: Capitol North/Center
Chair: Robert B. Schnabel, University of Colorado, Boulder

At the 1997 MPS Symposium in Lausanne, there were two talks on secant/quasi-Newton methods. Just 25 years ago, this area was as hot as interior point methods have been recently. What caused this decline in activity? In my talk, I will use that question to motivate the evolution of quasi-Newton methods since the seminal paper of Bill Davidon was brought to light by Roger Fletcher and Mike Powell.

I promise not to offer yet another explanation of the superiority of BFGS over DFP or to bore the audience with arcane historical detail. I will offer some speculation on current directions in NLP.

John E. Dennis
Noah Harding Professor and CAAM Department Chair
Rice University

OP99 Home


Program Updates

Speaker Index




MMD, 12/17/98