Closed tschm closed 1 year ago
@phschiele
The Markowitz version did not work with all variables either up or down. At least one variable has to be free after finding the first turning point
For now I have implemented Lopez de Prado (Bailey) and the Markowitz code. They are both children of the CLAUX class. The class exposes you to the mean, upper and lower bounds, covariance, a logger, a tolerance and stores the turning points. I would aim for some common API between the solvers...
https://github.com/phschiele/PyCLA