1
$\begingroup$

A draft research paper claims that $Q(p)=1-p_1 p_2 p_3 p_4 - p_2 p_3 p_6 p_7-p_1p_2$ is multilinear where $p_i = \mathbb P(e_i)$ and $e_i$ is a basic event of a component to fail.

I have learnt in LP course that even a function $Q_2(p)=p_1 p_2$ is quadratic when $p_1,p_2 \in \{0,1\}$, not LP. I find the part "linear" confusing in the word "multilinear". The claim is so that $Q(p)$ is a multilinear function. What does this "multilinear" mean? Is this quadratic constraint $Q_2(p)$ multilinear?

$\endgroup$
2
$\begingroup$

Simply, a function with several variables is called multilinear function if it's linear for each variable.

$\endgroup$
7
  • $\begingroup$ I agree. So the function given is not multilinear. $\endgroup$
    – GEdgar
    Mar 11 '13 at 21:33
  • $\begingroup$ Suppose $p_i\in[0,1]$. Now is the function $e^{p_1}$ linear where $e$ is the Euler number? What about $p_1 p_2$? Some examples and counter-examples could help... $\endgroup$
    – hhh
    Mar 11 '13 at 21:34
  • $\begingroup$ $Q$ is not multilinear but $Q_2$ it is so we say also $Q_2$ is bilinear. $\endgroup$
    – user63181
    Mar 11 '13 at 21:36
  • 1
    $\begingroup$ Because there's the $1$, so you can check for example that $Q(p_1+p'_1,p_2,p_3,p_4)\neq Q(p_1,p_2,p_3,p_4)+Q(p'_1,p_2,p_3,p_4)$. $\endgroup$
    – user63181
    Mar 11 '13 at 21:40
  • 1
    $\begingroup$ I'm sorry.en.wikipedia.org/wiki/Multilinear_map $\endgroup$
    – user63181
    Mar 11 '13 at 21:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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