Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I am interested in computing the maximum of the following function

$E[ ((XB) - E(XB))^TR ] $ subject to the constraint $E(XB)=1$ where $X$ is a $n\times m$ random matrix, $B$ is a $m \times 1$ variable vector, and $R$ is a $n\times 1$ random vector. $T$ is the transpose operator.

I would like to find the value of $B$ that maximizes this function.What are my options in terms of solutions to this problem?

Thanks

share|improve this question

1 Answer 1

up vote 0 down vote accepted

Expected value is linear, so if $X$ is the only random quantity in the expression $E[((XB) - E(XB))^T R] = 0$.

EDIT: with the question changed to make $R$ random as well, the objective becomes $E(B^T X^T R) - E(XB)^T E(R) = B^T C = C^T B$ where $C = E(X^T R) - E(X^T) E(R)$, while the constraint is $E(X) B = 1$ (I assume that's a vector of all $1$'s?).

Case 1: If $1$ is not in the column space of the matrix $E(X)$, the problem is infeasible: there are no solutions to the constraint.

Case 2: If $1$ is in the column space of $E(X)$ and $C^T$ is in the row space of $E(X)$, the objective is constant on the solution space.
Any solution of the constraint maximizes it.

Case 3: If $1$ is in the column space of $E(X)$ and $C^T$ is not in the row space of $E(X)$, the objective is unbounded. You can take a vector $V$ such that $E(X) V = 0$ and $C^T V > 0$, and by adding a suitable multiple of $V$ to any solution of $E(X) B = 1$ you get a solution with $C^T X$ arbitrarily large.

share|improve this answer
    
Oops I apologize R is a $n \times 1$ random vector, so the function is $E(B^TX^TR) - E(XB)^TE(R)$ subject to $E(XB)=1$ –  peterdevi Aug 30 '12 at 20:02

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.