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.

Consider vectors $v_i \in \mathbb{R}^n$, $z_i \in \mathbb{R}^m$, $i = 1,2,\ldots,N$, and matrices $X$ (positive definite), $F$, $G$ (of appropriate dimensions).

Consider $\alpha_i \in \mathbb{R}_{[0,1]}$, such that $\sum_{i=1}^N \alpha_i = 1$ (see the as the coefficients of a convex combination).

Can the inequality

$$ \sum_{i, j \neq i}^N \alpha_i \alpha_j v_i^\top X \left( F v_j + G z_j \right) \leq 0 $$

be made independent from the $\alpha_i$s?

Note that this is possible for $N=2$ because $\alpha_1 \alpha_2 = \alpha_1(1-\alpha_1)$ and it can be simplified.

share|improve this question
    
At least: is the condition $$ \sum_{i, j \neq i}^N v_i^\top X \left( F v_j + G z_j \right) \leq 0 $$ sufficient? –  Adam Apr 30 '12 at 7:42
    
We may want to think about the problem as $\max_{\alpha_1,...,\alpha_N} (...) \leq 0$. Can we get something interesting by taking the partial derivatives $\partial/\partial \alpha_i$ (and imposing them to be $0$)? –  Adam Apr 30 '12 at 11:56

Your Answer

 
discard

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

Browse other questions tagged or ask your own question.