No label defined (Q3738941)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 3972651
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 3972651 |
Statements
1986
0 references
comparison of algorithms
0 references
active-set logics
0 references
quasi-Newton
0 references
steepest- descent
0 references
dual-violator
0 references
worst-violator
0 references
0.8535970449447632
0 references
0.8048064708709717
0 references
0.7878432869911194
0 references
0.7762300372123718
0 references
The effectiveness of four active-set logics, Kelley's dual-violator rule, Sargent's worst-violator rule, Goldfarb's rule and Das' logic has been studied for three optimization processes of Newton, quasi-Newton and steepest-descent types. The dual-violator and worst-violator rules attempt to choose the active set as small as possible at some risk of infeasibility, while the other two logics drop constraints in a delicate and quite conservative manner which guarantees feasibility, to alleviate the cycling phenomenon often met with least-constrained strategies. An enumerative combinational logic, inefficient for large numbers of constraints, is also evaluated as a yardstick. In calculation for 24 quadratic test problems with linear constraints used in the present study, it is found that where relatively few constraints are active at the solution, the dual-violator and worst- violator rules performed better than or as well as the Goldfarb and Das logic, and in other cases where many constraints are active at the solution the Goldfarb and Das logics tend to perform comparatively better.
0 references