I try to use IPOPT to minimize a linear function under some bound constraints. According to the doc the following code should work:
real[int] a = [-1.0,-1.2];
real[int] UB = [1.0,1.0];
real[int] LB = [0.0,0.0];
real[int] x0 = [0.5,0.5];
However, I get a segmentation fault. I don’t know where the error comes from. Can you please indicate what I am doing wrong? Thank you.
I’ve never used IPOPT. But I’ve seen that there is some examples in examples/plugin. For instance IpoptMinSurf.edp. It looks like the first arguments should be a func, not a real[int].
In the documentation it says that the arguments determining the function can be of the form [A,b] when the function to be optimized is x^TAx+b^Tx (basic quadratic function), or a real[int] when the function is simply a linear function. Supposedly, for these basic cases, more efficient algorithms can be used.
I managed to make the algorithm work by supplying [A,b] with A=0. However, I was wondering why the version present in the documentation does not work.