Is there a general rule for how to write high order polynomials in matrix form?

For example a linear combination of parameters:

$$a_1 x_1+a_2 x_2+a_3 x_3 + \cdots+ a_n x_n$$

Can be written as

$$\sum^n_{i=1} a_i x_i = \vec{a}^T\vec{x} $$

Second order forms are given by

$$ (a_1 x_1+a_2 x_2+a_3 x_3 + \cdots+ a_n x_n)^2 = \vec{x}^T {\mathbf A} \vec{x}$$

Which ensures all combinations of second order terms. What about the higher orders? i.e.

$$(a_1 x_1+a_2 x_2+a_3 x_3 + \dots +a_n x_n)^k$$

What forms ensure all combinations of terms. Is there a general rule to this? Does it have a name?

  • $\begingroup$ You might be interested in Einstein notation. $\endgroup$ – Luca Bressan Sep 15 '16 at 14:51
  • 1
    $\begingroup$ Notice that $(\vec{a}\otimes\ldots\otimes \vec{a})^t(\vec{x}\otimes\ldots\otimes \vec{x})=(\vec{a}^t\vec{x})^k$, where $\otimes$ means the kronecker product of the column matrices $\vec{a}_{n\times 1},\vec{x}_{n\times 1}$. $\endgroup$ – Daniel Sep 15 '16 at 16:05

Linear Form

$$ \begin{pmatrix} a & b \\ \end{pmatrix} \begin{pmatrix} x \\ y \end{pmatrix}= ax+by$$

Quadratic Form

$$ \begin{pmatrix} x & y \end{pmatrix} \begin{pmatrix} a & b \\ b & c \end{pmatrix} \begin{pmatrix} x \\ y \end{pmatrix}= ax^2+2bxy+cy^2$$

Cubic Form

$$ \begin{pmatrix} x & y \end{pmatrix} \begin{pmatrix} ax+by & bx+cy \\ bx+cy & cx+dy \end{pmatrix} \begin{pmatrix} x \\ y \end{pmatrix}= ax^3+3bx^2y+3cxy^2+dy^3$$

Quartic Form

$$ \begin{pmatrix} x & y \end{pmatrix} \begin{pmatrix} ax^2+2bxy+cy^2 & bx^2+2cxy+dy^2 \\ bx^2+2cxy+dy^2 & cx^2+2dxy+ey^2 \end{pmatrix} \begin{pmatrix} x \\ y \end{pmatrix}$$

$m$-tuple Form

$$ \binom{m}{i_{1},i_{2}, \ldots , i_{n}} a_{i_{1} i_{2} \ldots i_{n}} x_{1}^{i_{1}} x_{2}^{i_{2}} \ldots x_{n}^{i_{n}}$$

where $\boldsymbol{x} \in \mathbb{R}^{n}$ and $i_{1}+i_{2}+\ldots+i_{n}=m$

| cite | improve this answer | |

I like what has been said already by @Daniel and @Ng Chung Tak but let me rephrase their answers.

We'll consider tensor products of the 1-form $a=(a_1,\ldots a_n)\in V^*$. A bilinear form $b=a\otimes a$ (which we'll call a "second order form" when supplied with the same $x\in V$ in both argument slots) can indeed be represented by a matrix equal to the outer product of the coordinates of $a$. A multilinear form $m=\otimes^ka$ would be represented by a multidimensional array.

Alternatively, you can look at this as an application of Multinomial theorem and multi-index notation

$$ (a_1x_1+\cdots+a_nx_n)^k = \sum_{|\alpha|=k}{k \choose \alpha}a^\alpha x^\alpha $$

| cite | improve this answer | |

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.