--- title: "Spacekime Analytics (Time Complexity and Inferential Uncertainty)" subtitle: "Coordinate Transformations, Covariance, Contravariance, and Invariance" author: "SOCR Team (Ivo Dinov)" date: "`r format(Sys.time(),'%m/%d/%Y')`" output: html_document: theme: spacelab highlight: tango includes: before_body: TCIU_header.html toc: true number_sections: true toc_depth: 2 toc_float: collapsed: false smooth_scroll: true code_folding: hide --- ```{r setup, include=FALSE, echo=FALSE} knitr::opts_chunk$set(echo = TRUE, warings = FALSE, error = TRUE) ``` # Coordinate Transformations Given complex vector space $V$, the set of all linear functions $L:V\to \mathbb{C}$ forms the dual space to $V$, which is also a vector space denoted by $V^*$. When $V$ is finite dimensional, $V^*$ has the same dimension as $V$. The pairing of vector spaces to their duals requires disambiguation of elements in $V$, called *contravariant vectors or kets*, and operatorsin $V^*$, called *dual vectors, covectors, covariant vectors, or bras*. *Covariant components* transform synergistically (“co”) in the same way as the basis vectors, whereas *contravariant components* transform anti-synergistically (“contra”) in the opposite way to basis vectors. *Transformation of basis vectors* refers to mapping of the basis vectors in the original coordinate system into new basis vectors. *Transformation of vector components* refers to the change in the vector components (coefficients) relative to two different sets of coordinate axes. $$\underbrace{(New\ basis\ vectors,\ ν')=(Transform\ matrix,\ T)\times (Original\ basis\ vectors,\ ν)}_{base\ vectors\ covary}\ ,$$ $$\underbrace{(Vector\ components\ in\ the\ new\ coordinate\ system,\ u_{ν'})=\\ (Inverse\ Transpose\ Transform\ matrix,\ (T^t)^{-1})\times (vector\ components\ in\ the\ original\ coordinate\ system,\ u_ν}_{base\ vectors\ covary}\ ,$$ Denote the new basis vector $v' = \begin{pmatrix} v_{1}' \\ v_{2}' \end{pmatrix}$, and the original basis vectors $v = \begin{pmatrix} v_{1} \\ v_{2} \end{pmatrix}$. Denote the vector components in the new coordinate system $u_{v^\prime} = \begin{pmatrix} u^{1\prime} \\ u^{2\prime} \end{pmatrix}$, and the vector in the old coordinate system $u_v = \begin{pmatrix} u^{1} \\ u^{2} \end{pmatrix}$. Any given (fixed) vector $\alpha\in \mathbb{R}^2$ would have unique representations in either the new ($v'$) or the old ($v$) coordinate systems $$\alpha \equiv \left \langle \begin{pmatrix} u^{1\prime}\\u^{2\prime} \end{pmatrix} | \begin{pmatrix} v_{1}^\prime \\ v_{2}^\prime \end{pmatrix} \right \rangle = \underbrace{\begin{pmatrix} u^{1\prime} & u^{2\prime} \end{pmatrix}}_{new\ components} \begin{pmatrix} v_{1}^\prime \\ v_{2}^\prime \end{pmatrix} \equiv \begin{pmatrix} u^{1} & u^{2} \end{pmatrix} \underbrace{\begin{pmatrix} v_{1} \\ v_{2} \end{pmatrix}}_{old\ basis}\ .$$ Then, there exists a *linear mapping* $T$ transforming the old basis into the new basis, $T:v\to v'$, $$\underbrace{\begin{pmatrix} v_1' \\ v_2' \end{pmatrix}}_{v'} = T \underbrace{\begin{pmatrix} v_1 \\ v_2 \end{pmatrix}}_{v}\ .$$ $$\alpha = \left \langle \begin{pmatrix} u^{1\prime}\\u^{2\prime} \end{pmatrix} | \begin{pmatrix} v_{1}^\prime \\ v_{2}^\prime \end{pmatrix} \right \rangle = \left \langle\begin{pmatrix} u^{1\prime}\\u^{2\prime} \end{pmatrix} |\ T \begin{pmatrix} v_{1} \\ v_{2} \end{pmatrix}\right \rangle = \left \langle T^t \begin{pmatrix} u^{1\prime}\\u^{2\prime} \end{pmatrix} | \begin{pmatrix} v_{1} \\ v_{2} \end{pmatrix}\right \rangle = \left \langle\begin{pmatrix} u^{1}\\u^{2} \end{pmatrix} | \begin{pmatrix} v_{1} \\ v_{2} \end{pmatrix}\right \rangle\ .$$ Due to the uniqueness of the coordinate components, $u_{v} = \begin{pmatrix} u^{1} \\ u^{2} \end{pmatrix}$ and $u_{v^\prime} = \begin{pmatrix} u^{1\prime} \\ u^{2\prime} \end{pmatrix}$, $$\underbrace{\begin{pmatrix} u^{1}\\u^{2} \end{pmatrix}}_{u_{v}} = T^t \underbrace{\begin{pmatrix} u^{1\prime}\\u^{2\prime} \end{pmatrix}}_{u_{v'}}\ .$$ The transposed basis transformation matrix $$T^t = \begin{bmatrix} T_{1}^{1} & T_{2}^{1} \\ T_{1}^{2} & T_{2}^{2} \end{bmatrix}$$ may be used to map the coefficient *components* of the vector $\alpha = u^1v_1 + u^2v_2$ from the old coordinate system, $v$, to components in the new basis, $v^\prime$: $$u_v = T^tu_{v\prime}\ \ \Longleftrightarrow \ \ u_{v^\prime} = \left(T^t\right)^{-1}u_v\ .$$ # Example Here is one example of a change of basis transformation $$\left| \begin{array}{cc} v_1' = v_1 + 2 v_2 \\ v_2' = v_1 + 3 v_2 \end{array} \right.\ .$$ $$T\equiv \begin{pmatrix} T_1^1 & T_1^2 \\ T_2^1 & T_2^2 \end{pmatrix}= \begin{pmatrix} 1 & 2 \\ 1 & 3 \end{pmatrix}\ , \ \ T^{-1} = \begin{pmatrix} 3 & -2 \\ -1 & 1 \end{pmatrix}\ ,$$ $$T^t\equiv \begin{pmatrix} T_1^1 & T_2^1 \\ T_1^2 & T_2^2 \end{pmatrix}= \begin{pmatrix} 1 & 1 \\ 2 & 3 \end{pmatrix}\ , \ \ (T^t)^{-1} = \begin{pmatrix} 3 & -1 \\ -2 & 1 \end{pmatrix}\ ,$$ $$ν'=Tν\ \ , \ \ v=T^{-1}v' \ .$$ Suppose the vector $u_ν=u^1 ν_1+u^2 ν_2$ in the old system and in the new coordinate system, $u_{ν'} = u_{ν'}^1 ν_1' + u_{ν'}^2 ν_2'$. then, $$u_ν=u^1 ν_1+u^2 ν_2 = \left\langle \begin{pmatrix} u^1 \\ u^2 \end{pmatrix} | \begin{pmatrix} v_1 \\ v_2 \end{pmatrix} \right\rangle= u^1 (3ν_1' - 2ν_2') + u^2 (-ν_1' + ν_2')=\cdots =\\ \underbrace{(3u^1-u^2)}_{u_{ν'}^1} ν_1' + \underbrace{(-2u^1 + u^2)}_{u_{ν'}^2} v_2' = u_{ν'}\ .$$