how to use the objective function minimize te in quadprog

I would like to make sure if this is correct

I have a set of initial weights [10x1] - w0 covariance matrix [10x10] - myCov

f = -w0'*myCov

[solution, fval, exitflag,output] = quadProg(mycovNew.data, f, A_le,b_le,Aeq,beq,zeros(n,1),[],w0, options)

Is this the correct syntax if I want to minimize TE with respect to initial set of weights.

Thanks!

[Matt J]

What's "TE"? Also, in one place you write myCov and later mycovNew.data. Are they the same thing?

[hdg D]

apologize te = tracking error and mycovNew = myCov Thanks, The objective is to minimize (w-w0)'myCov(w-w0)

Answer 1

[Matt J]

It's the correct syntax if the objective you're minimizing is

 w.'*myCov*w/2 - w0'*myCov*w

[Matt J]

The objective is to minimize (w-w0)'myCov(w-w0)

Yes. You have it correct.

[hdg D]

so the way i am using quadprog is correct then if my objective function is minimize (w-w0)'myCov(w-w0) and this is my entry for quadprog f = -w0'*myCov

[solution, fval, exitflag,output] = quadProg(myCov.data, f, A_le,b_le,Aeq,beq,zeros(n,1),[],w0, options)

Thanks a lot for your help Matt

[Matt J]

You're welcome. Be sure to close the Question (by Accept-clicking), if you have the Answer you need.

[R H]

Hi, to minimize (w-w0)'myCov(w-w0), don't you need an additional term: w0'*myCov*w0?

[Matt J]

The additional term is an additive constant, and therefore including/excluding it doesn't affect the optimization.