logo

Operations on Differential Forms: Sum and Wedge Product 📂Geometry

Operations on Differential Forms: Sum and Wedge Product

Definition1

Sum ++

Suppose ω=IaIdxI,φ=IbIdxI\omega = \sum\limits_{I} a_{I} dx_{I}, \varphi = \sum\limits_{I} b_{I} dx_{I} is in kk notation. Then, the sum of these two is defined as follows.

ω+φ:=I(aI+bI)dxI \omega + \varphi := \sum\limits_{I}\left( a_{I} + b_{I} \right)dx_{I}

Wedge Product \wedge

Let ω=IaIdxI\omega = \sum\limits_{I} a_{I}dx_{I}, φ=JbJdxJ\varphi = \sum\limits_{J} b_{J}dx_{J} be in kk notation and ss notation, respectively. Then, the wedge product of these two is defined as:

ωφ=I,JaIbJdxIdxJ \omega \wedge \varphi = \sum \limits_{I,J} a_{I}b_{J} dx_{I} \wedge dx_{J}

Explanation

Let MM be a nn-dimensional differential manifold. (If the concept of differential manifold is challenging, think of it as M=RnM=\mathbb{R}^{n}.) The wedge product \wedge turns two elements φ1,φ2\varphi_{1}, \varphi_{2} of TpMT_{p}^{\ast}M into an alternating function φ1φ2\varphi_{1} \wedge \varphi_{2}. Moreover, the wedge product itself is an alternating function. That is, for,

φiφj=φjφiandφiφi=0 \varphi_{i} \wedge \varphi_{j} = - \varphi_{j} \wedge \varphi_{i} \quad \text{and} \quad \varphi_{i} \wedge \varphi_{i} = 0

it is easy to demonstrate according to the definition.

(φ1φ2)(v1,v2)= det[φi(vj)]= φ1(v1)φ1(v2)φ2(v1)φ2(v2)= φ2(v1)φ2(vL)φ1(v1)φ1(v2)= (φ2φ1)(v1,v1) \begin{align*} \left(\varphi_{1} \wedge \varphi_{2}\right)\left(v_{1}, v_{2}\right) =&\ \det \left[\varphi_{i}\left(v_{j}\right)\right] \\ =&\ \begin{vmatrix} \varphi_{1}\left(v_{1}\right) & \varphi_{1}\left(v_{2}\right) \\ \varphi_{2}\left(v_{1}\right) & \varphi_{2}\left(v_{2}\right) \end{vmatrix} \\ =&\ -\begin{vmatrix} \varphi_{2}\left(v_{1}\right) & \varphi_{2}\left(v_{L}\right) \\ \varphi_{1}\left(v_{1}\right) & \varphi_{1}\left(v_{2}\right) \end{vmatrix} \\ =&\ -\left(\varphi_{2} \wedge \varphi_{1}\right)\left(v_{1}, v_{1}\right) \\ \end{align*}

(φ1,φ1)(v1,v2)= det[φ1(vi)]= φ1(v1)φ1(v2)φ1(v1)φ1(v2)= 0 \begin{align*} \left(\varphi_{1}, \varphi_{1}\right)\left(v_{1}, v_{2}\right) =&\ \det \left[\varphi_{1}\left(v_{i}\right)\right] \\ =&\ \begin{vmatrix} \varphi_{1}\left(v_{1}\right) & \varphi_{1}\left(v_{2}\right) \\ \varphi_{1}\left(v_{1}\right) & \varphi_{1}\left(v_{2}\right) \end{vmatrix} \\ =&\ 0 \end{align*}

By definition, the wedge product of a kk notation and a ss notation is a k+sk+s notation.

Moreover, while dxidxi=0dx_{i} \wedge dx_{i} = 0 for all ii, it is not the case that ωω=0\omega \wedge \omega = 0 always holds. For example, if ω=x1dx1dx2+x22dx3dx4\omega = x_{1}dx_{1}\wedge dx_{2} + x2_{2} dx_{3} \wedge dx_{4},

ωω=2x1x2dx1dx2dx3dx4 \omega \wedge \omega = 2x_{1}x_{2} dx_{1} \wedge dx_{2} \wedge dx_{3} \wedge dx_{4}

Example

Let ω=x1dx1+x2dx2\omega = x_{1}dx_{1} + x_{2}dx_{2} be a first-order form of R3\mathbb{R}^{3}, and φ=x1dx1dx2+x2dx1dx3+dx2dx3\varphi = x_{1}dx_{1}\wedge dx_{2} + x_{2} dx_{1}\wedge dx_{3} + dx_{2} \wedge dx_{3} be a second-order form of R3\mathbb{R}^{3}. Then,

ωφ= (x1dx1+x2dx2)(x1dx1dx2+x2dx1dx3+dx2dx3)= x12dx1dx1dx2+x1x2dx1dx1dx3+x1dx1dx2dx3+x1x2dx2dx1dx2+x22dx2dx1dx3+x2dx2dx2dx3= (x1x22)dx1dx2dx3 \begin{align*} \omega \wedge \varphi =&\ (x_{1}dx_{1} + x_{2}dx_{2}) \wedge (x_{1}dx_{1}\wedge dx_{2} + x_{2} dx_{1}\wedge dx_{3} + dx_{2} \wedge dx_{3}) \\ =&\ x_{1}^{2} \cancel{dx_{1} \wedge dx_{1}\wedge dx_{2}} + x_{1}x_{2}\cancel{dx_{1} \wedge dx_{1}\wedge dx_{3}} + x_{1}dx_{1} \wedge dx_{2} \wedge dx_{3} \\ & + \cancel{x_{1}x_{2}dx_{2} \wedge dx_{1}\wedge dx_{2}} + x_{2}^{2} dx_{2} \wedge dx_{1}\wedge dx_{3} + x_{2}\cancel{dx_{2} \wedge dx_{2} \wedge dx_{3}} \\ =&\ (x_{1}-x_{2}^{2}) dx_{1} \wedge dx_{2} \wedge dx_{3} \end{align*}

Properties

Let ω\omega be a kk notation, φ\varphi be a ss notation, and θ\theta be a rr notation. Then,

(ωφ)θ= ω(φθ)(ωφ)= (1)ks(φω)ω(φ+θ)= ωφ+ωθ,if r=s \begin{align} (\omega \wedge \varphi) \wedge \theta =&\ \omega \wedge (\varphi \wedge \theta) \\ (\omega \wedge \varphi) =&\ (-1)^{ks} (\varphi \wedge \omega) \\ \omega \wedge (\varphi + \theta) =&\ \omega \wedge \varphi + \omega \wedge \theta,\quad \text{if } r=s \end{align}

Proof

Proof(1)

Let ω=IaIdxI\omega = \sum\limits_{I} a_{I} dx_{I}, φ=JbJdxJ\varphi = \sum\limits_{J} b_{J} dx_{J}, ϑ=LcLdxL\vartheta = \sum\limits_{L} c_{L} dx_{L}.

(ωφ)ϑ= (IaIdxIJbJdxJ)LcLdxL= (I,JaIbJdxIdxJ)LcLdxLby definition of = I,J,L(aIbJ)cL(dxIdxJ)dxLby definition of = I,J,L(aIbJ)cLdxI(dxJdxL)by property of  \begin{align*} (\omega \wedge \varphi) \wedge \vartheta =&\ \left( \sum\limits_{I}a_{I}dx_{I} \wedge \sum\limits_{J}b_{J}dx_{J} \right) \wedge \sum\limits_{L} c_{L}dx_{L} \\ =&\ \left( \sum\limits_{I,J}a_{I}b_{J} dx_{I} \wedge dx_{J} \right) \wedge \sum\limits_{L} c_{L}dx_{L} & \text{by definition of } \wedge\\ =&\ \sum\limits_{I,J,L} (a_{I}b_{J})c_{L} (dx_{I} \wedge dx_{J}) \wedge dx_{L} & \text{by definition of } \wedge\\ =&\ \sum\limits_{I,J,L} (a_{I}b_{J})c_{L} dx_{I} \wedge (dx_{J} \wedge dx_{L}) & \text{by property of } \wedge\\ \end{align*}

Since the functions a,b,ca, b, c are real-valued, the associative law with respect to multiplication holds. Therefore,

(ωφ)ϑ= I,J,LaI(bJcL)dxI(dxJdxL)= IaIdxIJ,LbJcLdxJdxLby property of = ω(φϑ)by property of  \begin{align*} (\omega \wedge \varphi) \wedge \vartheta =&\ \sum\limits_{I,J,L} a_{I}(b_{J}c_{L}) dx_{I} \wedge (dx_{J} \wedge dx_{L}) \\ =&\ \sum\limits_{I} a_{I} dx_{I} \wedge \sum\limits_{J,L} b_{J}c_{L} dx_{J} \wedge dx_{L} & \text{by property of } \wedge\\ =&\ \omega \wedge ( \varphi \wedge \vartheta) & \text{by property of } \wedge\\ \end{align*}

Proof(2)

Let ω=IaIdxI\omega = \sum\limits_{I} a_{I} dx_{I}, φ=JbJdxJ\varphi = \sum\limits_{J} b_{J} dx_{J}.

ωφ= I,JaIbJdxIdxJ= I,JaIbJdxi1dxikdxj1dxjs= I,JaIbJ(1)kdxj1dxi1dxikdxj2dxjs= I,JaIbJ(1)k(1)kdxj1dxj2dxi1dxikdxj3dxjs = I,JaIbJ(1)ksdxj1dxj2dxjsdxi1dxik= (1)ksI,JaIbJdxIdxJ= (1)ksφω \begin{align*} \omega \wedge \varphi =&\ \sum\limits_{I,J} a_{I}b_{J} dx_{I} \wedge dx_{J} \\ =&\ \sum\limits_{I,J} a_{I}b_{J} dx_{i_{1}} \wedge \dots \wedge dx_{i_{k}} \wedge dx_{j_{1}} \wedge \dots \wedge dx_{j_{s}} \\ =&\ \sum\limits_{I,J} a_{I}b_{J} (-1)^{k} dx_{j_{1}} \wedge dx_{i_{1}} \wedge \dots \wedge dx_{i_{k}} \wedge dx_{j_{2}} \wedge \dots \wedge dx_{j_{s}} \\ =&\ \sum\limits_{I,J} a_{I}b_{J} (-1)^{k}(-1)^{k} dx_{j_{1}} \wedge dx_{j_{2}} \wedge dx_{i_{1}} \wedge \dots \wedge dx_{i_{k}} \wedge dx_{j_{3}} \wedge \dots \wedge dx_{j_{s}} \\ &\ \vdots \\ =&\ \sum\limits_{I,J} a_{I}b_{J} (-1)^{ks} dx_{j_{1}} \wedge dx_{j_{2}} \wedge \cdots \wedge dx_{j_{s}} \wedge dx_{i_{1}} \wedge \dots \wedge dx_{i_{k}}\\ =&\ (-1)^{ks} \sum\limits_{I,J} a_{I}b_{J} dx_{I} \wedge dx_{J} \\ =&\ (-1)^{ks} \varphi \wedge \omega \end{align*}

Proof(3)

Let ω=IaIdxI\omega = \sum\limits_{I} a_{I} dx_{I}, φ=JbJdxJ\varphi = \sum\limits_{J} b_{J} dx_{J}, ϑ=JcJdxJ\vartheta = \sum\limits_{J} c_{J} dx_{J}.

ω= ω(φ+ϑ)= ω(J(bJ+cJ)dxJ)by definition of += (IaIdxI)(J(bJ+cJ)dxJ)= I,JaI(bJ+cJ)dxIdxJby definition of  \begin{align*} \omega =&\ \omega \wedge (\varphi + \vartheta) \\ =&\ \omega \wedge \left( \sum\limits_{J} (b_{J} + c_{J}) dx_{J} \right) & \text{by definition of } + \\ =&\ \left( \sum\limits_{I} a_{I} dx_{I} \right)\wedge \left( \sum\limits_{J} (b_{J} + c_{J}) dx_{J} \right) \\ =&\ \sum\limits_{I,J} a_{I}(b_{J} + c_{J})dx_{I}\wedge dx_{J} & \text{by definition of } \wedge\\ \end{align*}

Since the functions a,b,ca, b, c are real-valued, the distributive law holds. Therefore,

ω= I,JaIbJdxIdxJ+I,JaIcJdxIdxJ= ωφ+ωϑby definition of  \begin{align*} \omega =&\ \sum\limits_{I,J} a_{I}b_{J}dx_{I}\wedge dx_{J} + \sum\limits_{I,J} a_{I}c_{J}dx_{I}\wedge dx_{J} \\ =&\ \omega \wedge \varphi + \omega \wedge \vartheta & \text{by definition of } \wedge\\ \end{align*}


  1. Manfredo P. Do Carmo, Differential Forms and Applications, p4-5 ↩︎