Solver with maximum number of non-zero inputs

[Antonio Pereira]

Hi everyone!
I am using solver to set up the best portfolio (which is just a set of weights) given an x number of stocks. For instance, I might have data for 100 stocks and with this data, solver will give me the optimal weights of each stock for my portfolio by maximizing a ratio that takes into account both expected returns and volatility (sharpe ratio). The problem that I am facing is related to the constraints: I am trying to set a maximum number of stocks (i.e., maximum number of non-zero values for weights) but solver keeps saying that it can't find any feasible solution. Important to note that there are in fact feasible solutions. Can anyone help me? By the way, as a constraint for the maximum number of stocks i was using a countif (non-zero) function which seems to work for a small number of stocks (I tried with only 5) but not for 100 as in my case.

Thanks in advance!

[Alf]

In a nutshell, Solver can not work with CountIf and If functions in the
model. This is because these are discontinuous functions. Solver is
not capable of determining a "finite difference" to help with a
derivative. Solver will often give up without warning at the first sign
of confusion.
Your model would have to be reworked without these functions.
= = = = = = =
Dana DeLouis

Still you could try to run your model using the Evolutionary solver engine as this is not as sensitive to discontinuous function. And I would also suggest you upload a file with your model.

Alf