Broyden updating, the good and the bad!
Andreas Griewank
Institut für Mathematik Humboldt Universität zu Berlin Unter den Linden 6 10099 Berlin Germany
Abstract
So far so good! We had an updating procedure (the 'ful secant method') that seemed to work provided that certain conditions of linear independence were satisfied, but the problem was that it did not work very well. In fact, it proved to be quite numerically unstable. (cf. [C. G. Broyden, On the discovery of the "good Broyden" method, Math. Program. 87, No. 2 (B), 209--213 (2000; Zbl 0970.90002)])