Linear programming

[ZDX]

Can Linear Programming Simplex Method VI solve the optimization problem with real value variables (x>=0 or x<0)? If yes, how?

[WinterIsHere!!!]

Hi, Use this for guidance

http://forums.ni.com/t5/LabVIEW/Linear-programming/m-p/3559171#M995873

http://zone.ni.com/reference/en-XX/help/371361H-01/gmath/linear_prog_simplex_method/

[ZDX]

Hi, Thanks for your reply. But in the guidance you provided, Linear programming Simplex Method VI is only used for nonnegitive variables, i.e., x>=0. I am wondering whether the VI can be used for both negative and nonnegative varibales or not. Any suggestion with more details? Thanks a lot.

[LocalDSP · Accepted solution]

Assuming your (potentially) negative variables have constraints (say -1000) you can offest them accordingly and adapt your input data correspondingly.

replace x0 >= -1000 with X0 = (x0 + 1000) >= 0

Don't use negative constraints unnecesseraly big as it may results in numerical issues.

[LocalDSP · Accepted solution]

I've created a quick example (Lv2014) that finds the minimax solution for three equations with two variables.

The solution has both variables negative so using the raw algorithm returns {0, 0}. If you increase the variables offset parameter, the actual solution { -4.86. -0.49} is found for any offset value >= 4.87.

I hope this works for you!

[ZDX]

Thanks for your help. It works for me!