答案1
也许是这样的?
\documentclass{article}
\usepackage{amsmath} % for 'align*' environment and '\stackrel' macro
\newlength\mylength
\settowidth{\mylength}{\scriptsize Lem.\ 1.1(ii)}
\begin{document}
\begin{align*}
(E\mathcal{V}_{i+1})^{\perp}
&\stackrel{\makebox[\mylength]{}}{=}
(EA^{-1}(E\mathcal{V}_i))^T \\
&\stackrel{\makebox[\mylength]{\scriptsize Lem.\ 1.1(i)}}{=}
E^{-T}(A^{-1}(E\mathcal{V}_i))^{\perp}\\
&\stackrel{\text{Lem.\ 1.1(ii)}}{=}
E^{-T}(A^T(E\mathcal{V}_i))^{\perp}\\
&\stackrel{\makebox[\mylength]{}}{=}E^{-T}
(A^T\widehat{\mathcal{W}}_{i+1})
=\widehat{\mathcal{W}}_{i+2}
\end{align*}
\end{document}
附录:如果您有很多这样的表达式,那么设置一个专用宏来表示专门的=
符号可能是一个好主意,例如按照以下代码行(它生成与上面显示的示例相同的输出)。如果符号上方没有任何内容需要排版=
,只需写入\myeq{}
即可获得正确的间距。
\documentclass{article}
\usepackage{amsmath} % for 'align*' environment and '\stackrel' macro
\newlength\mylength
\settowidth{\mylength}{\scriptsize Lem.\ 1.1(ii)} % width of longest label
\newcommand\myeq[1]{\stackrel{\makebox[\mylength]{\scriptsize #1}}{=}}
\begin{document}
\begin{align*}
(E\mathcal{V}_{i+1})^{\perp}
&\myeq{} (E A^{-1}(E\mathcal{V}_i))^T \\
&\myeq{Lem.\ 1.1(i)} E^{-T}(A^{-1}(E\mathcal{V}_i))^{\perp}\\
&\myeq{Lem.\ 1.1(ii)} E^{-T}(A^T(E\mathcal{V}_i))^{\perp}\\
&\myeq{} E^{-T} (A^T\widehat{\mathcal{W}}_{i+1})
=\widehat{\mathcal{W}}_{i+2}
\end{align*}
\end{document}
答案2
您可以使用eqparbox
;第一个参数\widerel
是任意标签(每个环境不同)。
但是,第二种解决方案似乎更好。当然,不建议对引理数进行硬编码,而\ref
应该使用。
\documentclass{article}
\usepackage{amsmath}
\usepackage{eqparbox}
\DeclareRobustCommand{\widerel}[2]{%
\mathrel{\eqmakebox[#1]{$\displaystyle#2$}}%
}
\begin{document}
\begin{align*}
(E\mathcal{V}_{i+1})^{\perp}
&\widerel{A}{=} (EA^{-1}(E\mathcal{V}_i))^T \\
&\widerel{A}{\overset{\text{Lem.\ 1.1(i)}}{=}} E^{-T}(A^{-1}(E\mathcal{V}_i))^{\perp}\\
&\widerel{A}{\overset{\text{Lem.\ 1.1(ii)}}{=}} E^{-T}(A^T(E\mathcal{V}_i))^{\perp}\\
&\widerel{A}{=} E^{-T}(A^T\widehat{\mathcal{W}}_{i+1})
=\widehat{\mathcal{W}}_{i+2}
\end{align*}
\begin{align*}
(E\mathcal{V}_{i+1})^{\perp}
&= (EA^{-1}(E\mathcal{V}_i))^T \\
&= E^{-T}(A^{-1}(E\mathcal{V}_i))^{\perp} &&\text{Lemma 1.1(i)}\\
&= E^{-T}(A^T(E\mathcal{V}_i))^{\perp} &&\text{Lemma 1.1(ii)}\\
&= E^{-T}(A^T\widehat{\mathcal{W}}_{i+1})
=\widehat{\mathcal{W}}_{i+2}
\end{align*}
\end{document}
答案3
这里有两种解决方案。第一种使用array
环境,将等号放在单独的列中。第二种解决方案是我推荐的,因为它更易于阅读;它将理由放在额外的列中,而不是等号上方。
\documentclass{article}
\usepackage{amsmath,array}
\newcommand\VV{{\mathcal V}}
\newcommand\WW{{\mathcal W}}
\begin{document}
\[\begin{array}{r@{}>{{}}c<{{}}@{}l}
(E\VV_{i+1})^{\perp}
&=&(EA^{-1}(E\VV_i))^T\\
&\stackrel{\text{Lem.1.1(i)}}{=}&E^{-T}(A^{-1}(E\VV_i))^{\perp}\\
&\stackrel{\text{Lem.1.1(ii)}}{=}&E^{-T}(A^T(E\VV_i))^{\perp}\\
&=&E^{-T}(A^T\widehat{\WW}_{i+1})\\
&=&\widehat{\WW}_{i+2}
\end{array}
\]
\end{document}
\documentclass{article}
\usepackage{amsmath}
\newcommand\VV{{\mathcal V}}
\newcommand\WW{{\mathcal W}}
\begin{document}
\begin{align*}
(E\VV_{i+1})^{\perp}
&=(EA^{-1}(E\VV_i))^T\\
&=E^{-T}(A^{-1}(E\VV_i))^{\perp} && \text{by Lem.~1.1(i)}\\
&=E^{-T}(A^T(E\VV_i))^{\perp} && \text{by Lem.~1.1(ii)}\\
&=E^{-T}(A^T\widehat{\WW}_{i+1})\\
&=\widehat{\WW}_{i+2}
\end{align*}
\end{document}