Index Notation

最新推荐文章于 2022-12-24 18:51:26 发布

Phantomlsh

最新推荐文章于 2022-12-24 18:51:26 发布

阅读量555

点赞数

分类专栏： Note 文章标签：张量 tensor 数学物理学

本文链接：https://blog.csdn.net/a1323933782/article/details/109542375

版权

Note 专栏收录该内容

7 篇文章 0 订阅

订阅专栏

Also known as Einstein Notation

Basic Rules

Dummy Index is any twice-repeated index in a single term, which is sumed over $1,2,3,\cdots,n$ . Typically, $n = 3$ . Can be replaced by other unused index with the same range.

$a_ib_i = \sum_{i=1}{n}a_ib_i$

Free Index is any non-repeated index in a single term, which can take only one value in a given term. Cannot be replaced by other index.
No index can exist more than twice in a given term.
In a equation using Einstein Notation, the free indices on both sides must match.

For Brackets, count the indices separately and take the maximum of each.

E.x.
$a_{ij}(b_i+c_j+d_k)$

$i$ twice in the first combined term, dummy index
$j$ twice in the second combined term, dummy index
$k$ once in the third combined term, free index

Non-Identities

$\begin{aligned} &1.\quad a_{ij}(x_i+y_j) \ne a_{ij}x_i + a_{ij}y_j \\ &2.\quad a_{ij}x_iy_j \ne a_{ij}x_jy_i \\ &3.\quad (a_{ij}+a_{ji})x_iy_j \ne 2a_{ij}x_iy_j \end{aligned}$

Identities

$\begin{aligned} &1.\quad a_{ij}(x_j+y_j) = a_{ij}x_j + a_{ij}y_j \\ &2.\quad a_{ij}x_iy_j = a_{ij}y_jx_i \\ &3.\quad a_{ij}x_ix_j = a_{ji}x_ix_j \\ &4.\quad (a_{ij}+a_{ji})x_ix_j = 2a_{ij}x_ix_j \\ &5.\quad (a_{ij}-a_{ji})x_ix_j = 0 \\ \end{aligned}$

Kronecker Delta

Define the symbol
$\delta_{ij} = \delta_i^j = \delta^{ij} = \left\{ \begin{matrix} 1, \quad i = j \\ 0, \quad i \ne j \end{matrix} \right.$

E.x.
$\delta_{ij}x_ix_j = x_ix_i$

Alternating Tensor

Define
$\epsilon_{ijk} = 0, \qquad\text{if any of }i,j,k\text{ are equal}\\ \epsilon_{123} = \epsilon_{231} = \epsilon_{312} = 1 \\ \epsilon_{321} = \epsilon_{213} = \epsilon_{132} = -1 \\$

Differentiation

$\frac{\partial x_i}{\partial x_j} = \delta_{ij}, \quad [\nabla f]_i = \frac{\partial f}{\partial x_i}, \quad \nabla\cdot\bm{u} = \frac{\partial u_j}{\partial x_j}, \quad [\nabla\times\bm{u}]_i = \epsilon_{ijk}\frac{\partial u_k}{\partial x_j}$