2

I have previously asked this question. To be self explanatory I would like to state it here once again:

I have the following code and I dont know how to align the two horizontal \Longleftrightarrows vertically higher than they are now. Ideally they need to be only vertically moved a bit up so that one of them is vertically between c_2 and c_1 and the other needs to be between c_3 and c_4.

\begin{align}
a(y)&=1-a(-y)\nonumber\\
&\Uparrow \nonumber\\
m_u=1/m_l \quad\quad\,\, c_2&=c_3 \quad\quad\,\,  \lambda_0=\lambda_1\nonumber\\
t_u=-t_l\Longleftrightarrow c_1&=c_4 \Longleftrightarrow \mu_0=\mu_1 \nonumber\\
&\Updownarrow \nonumber\\
f_1(y)&=f_0(-y)
\end{align}

Now I need however, brackets. Because without brackets it might mean different. I received an two very good answers for this question and one is even with brackets BUT it uses tikz-cd and arxiv doesnt accept it. See this question.. Therefore, I need to have the graph with brackets. Is there any way to do it without using tikz-cd and tikz?

2 Answers2

2

The result seems adequate; on the other hand, I think I'd have a hard time in trying to understand what the diagram means. But the document is yours. ;-)

\documentclass{article}
\usepackage{amsmath}

\begin{document}

\begin{equation}
\begin{array}{ccccc}
&&\makebox[0pt]{$\displaystyle a(y)=1-a(-y)$} \\[1ex]
&&\Uparrow \\
\left\{
  \!\begin{aligned}
  m_u&=1/m_l\\
  t_u&=-t_l
  \end{aligned}
\right\}
&
\Longleftrightarrow
&
\!\begin{aligned}
  c_2&=c_3 \\
  c_1&=c_4
\end{aligned}
&
\Longleftrightarrow
&
\left\{
  \!\begin{aligned}
  \lambda_0&=\lambda_1 \\
  \mu_0&=\mu_1
  \end{aligned}
\right\}
\\
&&\Updownarrow \\[1ex]
&&\makebox[0pt]{$\displaystyle f_1(y)=f_0(-y)$}
\end{array}
\end{equation}

\end{document}

enter image description here

egreg
  • 1,121,712
2

Here is a solution with the extensions to the cases environment from mathtools and a few adjustments with the makebox package:

\documentclass{article}

\usepackage{mathtools}
\usepackage{makebox} 

\begin{document}

\begin{align}%{2}
a(y)& =1-a(-y)\nonumber\\
& \makebox*{${}={}$}{$ \Uparrow $} \nonumber\\
\begin{rcases}\begin{cases} m_u =1/m_l \\t_u =-t_l \end{cases}\hskip-1.18em\end{rcases}\Longleftrightarrow\begin{cases} c_2\\c_1\end{cases}\hskip-1.18em & \begin{rcases} = c_3\\ = c_4 \end{rcases}
\Longleftrightarrow\begin{rcases}\begin{cases}\lambda_0=\lambda_1\\
 \mu_0=\mu_1\end{cases}\hskip-1.18em\end{rcases} \nonumber\\
&\makebox*{${}={}$} {$\Updownarrow$} \nonumber\\
f_1(y)& =f_0(-y)
\end{align}

\end{document} 

enter image description here

Bernard
  • 271,350
  • Missing brackets from the left and right can be corrected? – Seyhmus Güngören Jan 20 '15 at 22:30
  • @Seyhmus Güngören: See my updated answer. – Bernard Jan 20 '15 at 22:43
  • do you have any idea why the bracket size of the very left "cases" environment is smaller than the other two? – Seyhmus Güngören Jan 20 '15 at 22:56
  • I hadn't paid attention to this point. It comes from the order of the nestìng of the cases/rcases environments. It's fixed. Is is all right now? – Bernard Jan 20 '15 at 23:02
  • I think it is good enough. Do you have any idea why tikz seems still better than this? By the way this has also some nicer parts like the equations are aligned and that the equation number is just below everything. I think tikz has only nicer arrows, I would say. – Seyhmus Güngören Jan 20 '15 at 23:07
  • I have the document file \documentclass[12pt,draftcls,onecolumn]{IEEEtran} and when I use this, one of the brackets is becoming smaller! I am really schocked. Please let me know if you have any idea to resolve this. My final idea is that this answer is nicer and I want to use it but I need it in both of my documents. In the other document everything is fine. – Seyhmus Güngören Jan 20 '15 at 23:47
  • I don't really know. I used the default arrows from amsmath. actually I don't like so much default tikz-cd arrows, and I usually prefer to use the psmatrix environment from pstricks for commutative diagrams, as I feel I have more control on a lot of details, with a LaTeX-like syntax. – Bernard Jan 20 '15 at 23:49
  • I changed my code: it was enough to replace the cases environment with cases*. In addition, I don't have to make any horizontal adjustment any more, so the code is simpler. See the updated answer. – Bernard Jan 21 '15 at 00:06
  • Thank you very much once again. now it is free of problems and seems very nice. – Seyhmus Güngören Jan 21 '15 at 17:43