In coming versions of YALMIP, the operators > and < will most likely not be allowed in constraints. Hence, to prepare, you should replace occurrences of these operators with <= and >=.
The reason for this change is that strict operators cause confusion. The notion of a strict inequality has little support in practice, as almost all solvers approach the optimal solution from infeasibility, and thus easily ends up with non-strict feasibility, or even slight infeasibility. Along the same lines, a strict inequality has little support in theory, since a problem containing a strict inequality has no minimizer (only infinimizer). YALMIP has never treated strict and non-strict inequalities differently, but supported strict inequalities as a notational convenience.
If you need strictly feasible solutions, you should manually add a suitable margin to your non-strict constraint.
The next version of YALMIP will issue warnings when strict inequalities are encountered.
Update: YALMIP now warns when you use strict inequalities. This can be turned off by warning('off','YALMIP:strict'), although you of course should fix your model to begin with.
Unfortunately, SeDuMi has a bug which turns off all warnings, so the warning will typically not be shown by YALMIP once you have run SeDuMi.
Comments
Hello, recently I have an nonlinear semi-definite programming to be solved. I want to try pennon solver to have a shot. I saw the website said that a beta release could be asked. I am wondering if I can get it. Any help is appreciated, thanks.
You would have to ask the developers of pennon, www.penopt.com