Heinz Koeppl, Marc Hafner, et al.
BMC Bioinformatics
Inequality constrained minimization problems are often solved byconsidering a sequence of parameterized barrier functions. Each barrierfunction is approximately minimized and the relevant parameterssubsequently adjusted.It is common for the estimated solution to one barrier function problemto be used as a starting estimate for the next. However, this hasunfortunate repercussions for the standard Newton-like methods applied tothe barrier subproblem.In this note, we consider a class of alternative Newton methods whichattempt to avoid such difficulties.Such schemes have already proved of use in the Harwell Subroutine Libraryquadratic programming codes {\tt VE14} and {\tt VE19}. © 1994, Springer-Verlag Berlin Heidelberg. All rights reserved.
Heinz Koeppl, Marc Hafner, et al.
BMC Bioinformatics
A.R. Conn, Nick Gould, et al.
Mathematics of Computation
Kenneth L. Clarkson, K. Georg Hampel, et al.
VTC Spring 2007
Robert F. Gordon, Edward A. MacNair, et al.
WSC 1985