CODEWITHSUNDEEP
pythonoptimizationscipy-optimize-minimize
Mat
Mat

Reputation: 41

Are Scipy Optimize Bounds inclusive?

I am trying to minimize a complex objective function that has 2 decision variables

The variables have bounds as mentioned below:

0<= var1/var2 < Some_upper_bound

As per my understanding of bounds variable in the optimize.minimize() function, both upper and lower values are inclusive.

How do I create a bound such that one value is inclusive (0) and the other is exclusive (Some_upper_bound)?

Any help with this is much appreciated. Thanks in advance!

Upvotes: 0

Views: 665

Answers (1)

Erwin Kalvelagen
Erwin Kalvelagen

Reputation: 16724

No tool will do what you want. For a reason: the feasible region is no longer compact and the optimization problem is no longer well defined. Also, most solvers actually use a feasibility tolerance, so their constraint effectively becomes: x <= a + feastol. This is done because computations suffer from limited floating point precision. See: https://yetanothermathprogrammingconsultant.blogspot.com/2017/03/strict-inequalities-in-optimization.html.

Upvotes: 1

Related Questions