# Could not find a feasible initial point.

12 views (last 30 days)
Mehul Agrawal on 1 Jun 2016
Edited: Matt J on 1 Jun 2016
I am trying to solve an optimization problem using ga where I have some linear constraints (both equality and inequality). Catch is that I have a lot of variables (~20K variables). My problem always terminates with "Could not find a feasible initial point." How can I resolve that ?
Mehul Agrawal on 1 Jun 2016
To give some estimate of the code, my problem is a packing problem where variables are to lie between 0 and 1. The constraints are that some variables are less than a certain number (less than 1 but greater than 0). So, there should be plenty of feasible initial points (agreed that they may be sub optimal).
Matt J on 1 Jun 2016
Edited: Matt J on 1 Jun 2016
We need to see code - at the very least the part where you set up the constraint data and your call to ga().

John D'Errico on 1 Jun 2016
Edited: John D'Errico on 1 Jun 2016
"plenty" of feasible points is not relevant, because you have provided no evidence that this is true. I see no reason why you draw that conclusion. You have a problem in 20,000 dimensions. That is a huge search space. If you cannot give it a good hint of where to start, AND you have sufficiently problematic constraints, then expect problems.
I would note that if it is so trivial to find a feasible start point, then why did you not provide one?
My guess is that you have equality constraints that may actually preclude any solution, or they are sufficiently constraining that there are far fewer feasible solutions than you apparently expect. Again, 20,000 dimensions is a BIG search space.

Mehul Agrawal on 1 Jun 2016
The equality constraints that I have are like x1 = 0.5 (No fancy equation). The other constraints are like x2+ x3+ x4 = 1 where x2, x3, x4 are (0,1). [These are basically all the constraints, just expand to 20,000 variables]
I can say that there are plenty of solutions because this is being worked in a real company and everyday the workers pack the bag randomly while still not violating any constraint. The idea of this problem is just to find the best bag to manufacture and create a stability in structure.