You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+8-2Lines changed: 8 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -31,9 +31,15 @@ A work in progress.
31
31
* backtracking linesearch method
32
32
* exact linesearch method using [fmin](https://github.com/jacobwilliams/fmin) minimizer
33
33
* evaluate function at specified fixed points
34
-
* Has two options for variable bounds (`xlow<=x<=xupp`):
34
+
* Has four options for variable bounds (`xlow<=x<=xupp`):
35
35
* Ignore bounds
36
-
* Crude method: manually adjust `x` vector at each function evaluation so that `x = min(max(x,xlow),xupp)`.
36
+
* Scalar method: The Newton search direction is changed by checking each `x` component individually. If a component is on a bound, its value is set to the bound. The result is the new line search direction.
37
+
* Vector mode: The Newton search direction is not changed, but, the magnitude is walked back so that no variable violates their bounds.
38
+
* "Wall" mode: Similar to scalar mode, except after modifying the search direction vector, each line search step keeps any values constant that were put on the bounds.
39
+
* Has three options for computing the function vector norm:
0 commit comments