Simplex method in prime how to use the constrains maximise s.t. constraints.

[ftneek]

(11-10-2023 04:05 PM)Albert Chan Wrote:  I like the machine to take care of slack variables. Less typing, less errors.
Rename simplex as simplex1, I added simplex(a) wrapper that add slack stuff.
.
.
.
> simplex([-[2,4,22],-[3,2,20],-[4,5,40],[6,5,-1]])

I have to admit the negative/positive argument in the c row for min/max is a nice idea.

I agree it would be nice if slack/artificial variables aren't required to be entered by the user. My only concern with this is approach is now all constraints are assumed to be "<=". If an "=" constraint is present, an artificial variable should be added instead of a slack variable. Therefore like simplex_reduce(), this wrapper method would be limited to solving systems in the form A.x<=b (but with no restriction on b), where as my simplex program can be used for mixed constraints as long as appropriate variables are added. And eventually I would like get around to handling mixed constraints without the user needing to add any additional variables.

For this reason I think the name of the wrapper should change to simplex_red() to indicate some similarities with simplex_reduce(), or simplex_le() to indicate that it is only for use with "less than or equal to (<=)" constraints. Then it is more clear (to me) that it could be used as a substitute/replacement for simplex_reduce(). Thank you for the examples!