abstract: stuffs involved exterior algebra

外代数

wedge product

包含vector spaces 的直和,性质\((1,\text{a})\)以及如下性质叫做外代数(exterior algebra)

a). 双线性(bilinear) \[ \begin{align} \left(\varphi_{1}+\varphi_{2}\right) \wedge \psi &=\varphi_{1} \wedge \psi+\varphi_{2} \wedge \psi \\ \varphi \wedge(a \psi) &=a(\varphi \wedge \psi), \quad a \in \mathbb{R} \end{align} \]

b). associative \[ (\varphi \wedge \psi) \wedge \chi=\varphi \wedge(\psi \wedge \chi) \]

注意:对于有限个线性空间(或向量空间)一般认为直和就是直积,但是直积和张量积不同。

补充

:张量的形式,例:\((2,3)\)\(\gamma^{\rho \sigma}_{ijk}\) ; \((0,2)\)\(\gamma_{\mu\nu}\)\((n,m)\)\(\gamma_{\rho_1,\dots,\rho_m}^{\sigma_1,\ldots,\sigma_n}\)

勘误

\([\ ]\)代表对易子,如 \([A,B]=AB-BA\) , \(\{\ \}\) 代表反对易子 \(\{A,B\}=AB+BA\)

辨析:Cartesian product tensor product, direct product (Kronecker product)

Cartesian product: 如果\(A, B\) 是集合,\(a\in A \ \& \ b\in B\)。then Cartesian product is \[ C=A \times B=\{(a, b) \mid a \in A, b \in B\} \] Tensor product 两个vector space 的线性映射,而 kronecker product 作用于两个 matrixs

kronecker product

: If \(A\) is an \(m × n\) matrix and \(B\) is a \(p × q\) matrix, then the Kronecker product \(A \otimes B\) is the \(pm × qn\) block matrix: \[ \mathbf{A} \otimes \mathbf{B}=\left[\begin{array}{ccc} a_{11} \mathbf{B} & \cdots & a_{1 n} \mathbf{B} \\ \vdots & \ddots & \vdots \\ a_{m 1} \mathbf{B} & \cdots & a_{m n} \mathbf{B} \end{array}\right] \] tensor product:

If \(\left(e_{1}, \ldots, e_{n}\right)\) is a basis of \(V\) and \(\left(f_{1}, \ldots, f_{m}\right)\) is a basis of \(W\), then the set of elements \(e_{i} \otimes f_{j}\) where \(i=1, \ldots, n\) and \(f=1, \ldots, m\) forms a basis for \(V \otimes W .\) It is simple to see these span the space since for any \(v \otimes w\) we have \(v=\sum_{i} v_{i} e_{i}\) and \(w=\sum_{j} w_{j} f_{j}\) so that \[ v \otimes w=\left(\sum_{i} v_{i} e_{i}\right) \otimes\left(\sum_{j} w_{j} f_{j}\right)=\sum_{i, j} v_{i} w_{j} e_{i} \otimes f_{j} \] Given this, we see that the basis also spans linear superpositions of elements of the form \(v \otimes w\), thus general elements of \(V \otimes W\). With \(n \cdot m\) basis vectors, the dimensionality of \(V \otimes W\) is equal to the product of the dimensionalities of \(V\) and \(W\) : \[ \operatorname{dim}(V \otimes W)=\operatorname{dim}(V) \times \operatorname{dim}(W) \] Dimensions are multiplied (not added) in a tensor product. How do we construct operators that act in the vector space \(V \otimes W ?\) Let \(T\) be an operator in \(V\) and \(S\) be an operator in \(W\). In other words, \(T \in \mathcal{L}(V)\) and \(S \in \mathcal{L}(W)\). We can then construct an operator \(T \otimes S\) \[ T \otimes S \in \mathcal{L}(V \otimes W) \] defined to act as follows: \[ T \otimes S(v \otimes w) \equiv T v \otimes S w \] 张量概念:

\(V\) 是域 \(F\) 上的 \(n\) 维线性空间, \(q\)\(V\) 的张量积 \[ T^{q}(V)=V \otimes \cdots \otimes V \] 中任一元素称为 \(\boldsymbol{V}\) 上的一个 \(\boldsymbol{q}\) 秩反变张量 (q-contravariant tensor); \(p\)\(V^{*}\) 的张量积 \[ T_{p}(V)=V^{*} \otimes \cdots \otimes V^{*} \] 中任一元素称为 \(V\) 上的一个 \(p\) 秩协变张量 (p-covariant tensor); \(p\)\(V^{*}\)\(q\)\(V\) 的张量 积 \[ T_{p}^{q}(V)=V^{*} \otimes \cdots \otimes V^{*} \otimes V \otimes \cdots \otimes V \] 中任一元素称为 \(V\) 上的一个 \((p, q)\) 型张量, 也称为 \(V\) 上的一个 \(p\) 秩协变且 \(q\) 秩反变的混合张量 (\(p\)-covariant and \({q}\)-contravariant mixed tensor)。

多重线性张量代数

\(T^{\varphi}(V)\) 中任一张量 \(\boldsymbol{\alpha}\) 分别由上述两个基线性表出, 紧凑地写成

\[ \begin{align} \alpha &=a^{ {i_1}{i_2}\cdots i_{q} } \alpha_{i_{1} } \otimes \alpha_{i_{2} } \otimes \cdots \otimes \alpha_{i_{q} }\\ \alpha &=\tilde{a}^{j_{1} j_{2} \cdots j_{v} } \eta_{j_{1} } \otimes \eta_{j_{2} } \otimes \cdots \otimes \eta_{j_{q} } \tag{1.8} \end{align} \]

基于 \[ \begin{array}{ll} \alpha_{j}=b_{j}^{i} \eta_{i}, & j=1,2, \cdots, n \\ \alpha^{j}=a_{i}^{j} \eta_{i}, & j=1,2, \cdots, n \end{array} \]\(\boldsymbol{a}\) 张量可表示为: \[ \boldsymbol{\alpha}=a^{i_{1} i_{2} \cdots i_{q} }\left(b_{i_{1} }^{j_{1} } \eta_{j_{1} }\right) \otimes\left(b_{i_{2} }^{j_{2} } \eta_{j_{2} }\right) \otimes \cdots \otimes\left(b_{i_{q} }^{j_{q} } \eta_{j_{q} }\right) \]

\[ =b_{i_{1} }^{j_{1} } b_{i_{2} }^{j_{2} } \cdots b_{i_{q} }^{j_{q} } a^{ {i_{1} }{i_{2} \cdots i_{q} }} \eta_{j_{1} } \otimes \eta_{j_{2} } \otimes \cdots \otimes \eta_{j_{q} } \]

\(\left(b_{i_{2} }^{j_{2} } \eta_{j_{2} }\right) \otimes\left(b_{i_{q} }^{j_{q} } \eta_{j_{q} }\right)\) 因为 \(b_{i_{2} }^{j_{2} }\) 是常数可以直接提出前面,而 \(\eta_{j_{q} }\) 是 向量所以直接乘。

从式 (1.8) 得 \[ \tilde{a}^{j_{1} j_{2}\cdots{j}_{q} }=b_{i_{1} }^{j_{1} } b_{i_{2} }^{j_{2} } \cdots b_{i_{q} }^{j_{q} } a^{i_{1} i_{2} \cdots i_{q} } \] 注:楔积的解释

勘误

以下解释为错误解释

(不要轻信知乎或者其他论坛上的人的总结,要忠于教材)

\(f \in A_{k}(V), g \in A_{l}(V)\) ,定义他们的楔积 (wedge product) \(f \wedge g\) 为: \[ f \wedge g=\frac{1}{k ! l !} Alt(f \otimes g) \tag{1.9-old} \] 这里 \(Alt\) 表示反对称

正确的解释

\(f \in A_{k}(V), g \in A_{l}(V)\) ,定义他们的楔积 (wedge product) \(f \wedge g\) 为: \[ \begin{aligned} &\small\tau \wedge \sigma=\frac{(r+p) !}{r ! p !} A l t(\tau \otimes \sigma)\\ &\scriptsize=\frac{1}{r ! p !} \sum_{\gamma \in \mathcal{S}_{r+p} } \operatorname{sgn}(\gamma)\tau^{i_{\gamma(1)} \ldots i_{\gamma(r)} } \sigma^{i_{\gamma(r+1)} \ldots i_{\gamma(r+p)} } e_{i_{1} } \otimes e_{i_{2} } \otimes \cdots \otimes e_{i_{r+p} }\\ \end{aligned} \tag{1.9-new} \]

\(\Lambda^pV\) 的一个基: \[ \beta^{i_{1} } \wedge \beta^{i_{2} } \wedge \cdots \wedge \beta^{i_{p} }, \quad 1 \leqslant i_{1}<i_{2}<\cdots<i_{p} \leqslant n \] 于是有: \[ \begin{aligned} \operatorname{dim} \Lambda^{p} V &=\left(\begin{array}{l} n \\ p \end{array}\right) \\ \operatorname{dim} \Lambda V &=2^{n} \end{aligned} \] 注意:对于理论物理而言有时不需要过多注意基础理论的的证明,但是要会用这个结论去证明其他的东西。

任何 \(p- \text{from}\)\(\varphi \in \Lambda^pV\) 可以如下分解(表示)

\[ \varphi=\frac{1}{p !} \sum_{i_{1}, \ldots, i_{p} } \varphi_{i_{1} \cdots i_{p} } \beta^{i_{1} } \wedge \cdots \wedge \beta^{i_{p} } \] wedge product of \(p\)-form \(\varphi\) and a \(q\)-form \(\psi\) corresponds to the antisymmetrized tensor product up to a factor \((p+q) !\) : \[ \varphi \wedge \psi \leftrightarrow(p+q) ! \varphi_{\left[i_{1} \cdots i_{p}\right.} \psi_{\left.j_{1} \cdots j_{q}\right]} \tag{1.10} \] 方括号是指完全反对称。

problem 这里(1.9) 与式(1.10)有什么区别,为什么会出现不一样?

补充

反对称括号

The antisymmetrization of this tensor is defined by; \[ A l t\left(v_{1} \otimes \cdots \otimes v_{r}\right)=\frac{1}{r !} \sum_{\gamma \in \mathcal{S}_{r} } \operatorname{sgn}(\gamma) v_{\gamma(1)} \otimes \cdots \otimes v_{\gamma(r)} \] 张量计算 \(T^{q}(V)\) 可记成 \(T_{o}^{q}(V) ; T_{p}(V)\) 可记成 \(T_{p}^{\circ}(V)\) 。此外, 规定 \[ T_{0}^{0}(V)=F \] 由于 \(T_{p}^{q}(V)\) 是域 \(F\) 上的线性空间, 因此它有加法和纯量乘法运算。在取定的一个基下, 用坐标表示的张量的加法是把对应分量相加, 因此可以用下式来表示 \((p, q)\) 型张量的 加法: \[ (c+d)_{j_{1} j_{2} \cdots j_{p} }^{i_{1} i_{2} \cdots i_{q} }=c_{j,_{1} j_{2} \cdots j_{p} }^{i_{1} i_{2} \cdots i_{q} }+d_{j_{1} j_{2} \cdots j_{p} }^{i_{1} i_{2} \cdots i_{q} } \] 类似地, 可以用下式表示 \((p, q)\) 型张量的纯量乘法: \[ (k c)_{j_{1} j_{2} \cdots j_{p} }^{i_{1} i_{2} \cdots i_{q} }=k c_{j_{1} j_{2} \cdots j_{p} }^{i_{1} i_{2} \cdots i_{q} } \] 在同构的意义下 \[ T_{p}^{q}(V) \otimes T_{r}^{s}(V) \cong T_{p+r}^{q+s}(V) \] 设同构映射为 \(\psi\), 任给 \(\boldsymbol{f} \in T_{p}^{q}(V), \boldsymbol{g} \in T_{r}^{s}(V)\), 规定 \(\boldsymbol{f}\)\(\boldsymbol{g}\) 的乘积为 \[ \psi(\boldsymbol{f} \otimes \boldsymbol{g}) \in T_{\boldsymbol{p}+\boldsymbol{r} }^{\mu+s}(V) \]\(\boldsymbol{f}\)\(\boldsymbol{g}\) 的乘积仍记成 \(\boldsymbol{f} \otimes \boldsymbol{g}\) 。于是有了张量的乘积, 即 \(T_{P}^{q}(V)\) 中的张量与 \(T_{r}^{s}(V)\) 中的张 量的乘积为 \(T_{o+r}^{q+s}\) 中的张量。在 \(V\) 中取一个基, 分别得到 \(T_{p}^{q}(V), T_{r}^{s}(V), T_{o+r}^{q+s}(V)\) 的一个

Inner derivative

对于任何vector space \(V\) 内的vector \(v\), inner derivative \(i_v\) 可以降低一个degree \[ \begin{aligned} i_{v}: \Lambda^{p} V & \rightarrow \Lambda^{p-1} V \\ \varphi & \mapsto i_{v} \varphi, \quad v \in V \\ \left(i_{v} \varphi\right)\left(v_{1}, \ldots, v_{p-1}\right): &=\varphi\left(v, v_{1}, \ldots, v_{p-1}\right) \end{aligned} \] 如果 \(\varphi\)\(1-\text{form}\) 于是: \(i_{v} \varphi=\varphi(v) \in \mathbb{R}\)