I have a 2d PDE and NDSolve returns a solution that takes negative values but I want the function to be positive
f[x,y]>=0
everywhere. How do I include this simple constraint?
Asked
Active
Viewed 26 times
0
-
The sign of the solution will be defined by the initial and boundary conditions and the range over which you integrate. Therefore, you have to adjust these to get a solution with f>0. – Daniel Huber Mar 05 '21 at 10:35
-
is there no way of including this constraint? – VN23 Mar 05 '21 at 10:39
-
You are not free to specify this. It is given be the afore mentioned parameters and of course the equation. Another question would be, if you are searching these parameters. – Daniel Huber Mar 05 '21 at 10:44
-
its a physical problem, I understand its not a mathematically 100% well-posed PDE but I dont think NDSolve is so strict – VN23 Mar 05 '21 at 10:52
-
If physics requests f>0, then I think either the equations or inital/boundary conditions are wrong. And it is also possible that you have a problem with numerical stability. And of course the possibility exists, that you are on the brink to detect new physics. – Daniel Huber Mar 05 '21 at 10:56
-
2@VN23 Also we can't tell you nothing without PDE, boundary conditions and code you try to solve your problem with. – Alex Trounev Mar 05 '21 at 13:33