[Home]
[
News]
[ Events]
[ People]
[ Research]
[ Education]
[Visitor Info]
[UCSD Only]
[Admin]
Numerical Results for a Projected-Search Interior Method
by Philip E. Gill, Minxin Zhang
Abstract:
Numerical results are presented for a new interior-point method for constrained optimization that combines a shifted primal-dual interior-point method with a projected-search method for bound-constrained optimization. The method involves the computation of an approximate Newton direction for a primal-dual penalty-barrier function that incorporates shifts on both the primal and dual variables. The approximate Newton direction is used in conjunction with a new projected-search line-search algorithm that employs a flexible non-monotone quasi-Armijo line search to obtain an improved value of the penalty-barrier function. The beneficial effects of shifting both the primal and dual variables and the use of a projected search are illustrated by results from two methods that do not use projection. The results show that the all-shifted method is more efficient than the method that shifts only the primal variables. Moreover, the method using projection is more robust and requires substantially fewer iterations. In particular, the number of times that the search direction must be computed is significantly reduced.
UCSD-CCoM-22-03.pdf June 2022
News]
[ Events]
[ People]
[ Research]
[ Education]
[Visitor Info]
[UCSD Only]
[Admin]
Directors:
Randolph E. Bank
Philip E. Gill
Michael Holst
Administrative Contact:
[ Click to Send Email ]
Randolph E. Bank
Philip E. Gill
Michael Holst
Administrative Contact:
[ Click to Send Email ]
by Philip E. Gill, Minxin Zhang
Abstract:
Numerical results are presented for a new interior-point method for constrained optimization that combines a shifted primal-dual interior-point method with a projected-search method for bound-constrained optimization. The method involves the computation of an approximate Newton direction for a primal-dual penalty-barrier function that incorporates shifts on both the primal and dual variables. The approximate Newton direction is used in conjunction with a new projected-search line-search algorithm that employs a flexible non-monotone quasi-Armijo line search to obtain an improved value of the penalty-barrier function. The beneficial effects of shifting both the primal and dual variables and the use of a projected search are illustrated by results from two methods that do not use projection. The results show that the all-shifted method is more efficient than the method that shifts only the primal variables. Moreover, the method using projection is more robust and requires substantially fewer iterations. In particular, the number of times that the search direction must be computed is significantly reduced.
UCSD-CCoM-22-03.pdf June 2022