wilmott.com

Hello,I have set up my optimization problem to minimize variance, but I keep on getting the "wrong" weights, in the sense that they don't add up to 1. The assumption I am taking here is that there is no short -selling. The error message I get in Matlab is: The system of equality constraints is consistent.Removing the following dependent constraints before continuing: 2 3 4The equality constraints are dependent.Exiting: the constraints are overly stringent;Can someone please explain to me what I might be doing?I have attached my code. Note on the code: I have thought about making ub = zeros(4,1). However, then, that just allocates the weights to some number (28.9017)- all allocation is the same. I don't think this is correct. Thanks for your feedback!

The Aeq matrix does not make much sense to me at all.What constraints are you trying to represent? You need only one constraint: that the weights add up to one, and you don't have that. The coefficients for that constraint would be a row of ones, and the right hand side would be a scalar 1.

I think the code is a result of my doing various testing of inputs. How does one represent that the weights equal one? If I put a row of 1's for Aeq, I get an error message saying that it should be a matrix. If I make it then, a row matrix, with zeros below into a 4x4, then I get all weights to be zero. I don't think this is correct at all. Have I misunderstood your explanation? Thank you very much for your contribution.

maybe try Aeq= transpose(ones(4));orAeq= reshape(ones(4),[1 4]);

The first Aeq works when I change f to equal zeros(4,1).