3

I have a set of equations summarized in this way:

enter image description here

The code that produces this is the following:

\documentclass[12pt]{article}    
\usepackage{empheq}
\usepackage[left=2.5cm,top=2.5cm,right=2.5cm,bottom=2.5cm]{geometry}
\usepackage{amsmath}

\begin{document}

\newcommand*\widefbox[1]{\fbox{\hspace{2em}#1\hspace{2em}}}
\begin{subequations}
\begin{empheq}[box=\widefbox]{align}
%
\Aboxed{a = & b + c + d}
 \nonumber \\
& d = f + g
 \nonumber \\
& \qquad
f = m/4
 \nonumber \\
&
d = j + k
 \nonumber \\
& \qquad
j = n\cdot3/4
 \nonumber \\
& d = l + o
 \nonumber \\
\Aboxed{& d = m/4 + q + n\cdot 3/4 + k + l + 0 }
 \nonumber \\
\Aboxed{&c = p + q}
 \nonumber \\
& \qquad
p = h/2 + h/2
 \nonumber \\
\Aboxed{&b = r + s}
 \nonumber \\
\Aboxed{&t = u + w}
 \nonumber \\  \nonumber\\
\Aboxed{a = & b + c + d}
 \nonumber \\ \nonumber\\
\Aboxed{a = & zz + z'}
 \nonumber \\ \nonumber\\
\end{empheq}
\end{subequations}

\end{document}

I would like to produce something similar to this:

enter image description here

Is there a way to achieve those arrows, red boxes and circled numbers 1, 2 and 3 ?

Update:

Following @Ignasi's approach, when I try to apply this to the real example, I encounter quite a lot of difficulties, this is the nearest result I could achieve:

enter image description here

where:

1) The circled numbers and the arrows appear in the following page, instead of next to the equations (see image)

2) I could not manage to align the J[p] equations.

3) I could not manage to box the last equation,

4) Is there a way to caption this scheme?

Could you please help me to obtain this result?

This is the code to where I reached so far:

\documentclass[12pt]{article}    
\usepackage{empheq}
\usepackage[left=2.5cm,top=2.5cm,right=2.5cm,bottom=2.5cm]{geometry}
\usepackage{amsmath}
\usepackage{tikz}
\usetikzlibrary{tikzmark, positioning}

\begin{document}

\newcommand*\widefbox[1]{\fbox{\hspace{2em}#1\hspace{2em}}}

\begin{subequations}
\begin{empheq}[box=\widefbox]{align}
%
%
\tikzmark{1}\Aboxed{
E [\rho ] = & \underbrace{ T[\rho ] + V_{ee}[\rho ] }_{=\,\, F[\rho ]} + V_{ne}[\rho]
}
 \nonumber \\
& %
\tikzmark{2}
V_{ne}[\rho ] = \int \rho \left ( \mathbf{r} \right ) v_{\text{ext}} \left ( \mathbf{r} \right ) \mathrm{d}\mathbf {r}
 \nonumber \\
& \qquad \qquad
v_{\text{ext}}\left ( \mathbf{r}_{i} \right ) = - \sum_{A=1}^{M} \frac{Z_{A}}{r_{iA}}
\nonumber \\
&
\tikzmark{3}
V_{ne}[\rho ] =  \int -  \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}} \rho \left ( \mathbf{r}_{1} \right )  \mathrm{d}\mathbf {r}_{1}
 \nonumber \\
& \qquad \qquad
\rho (\mathbf{r}) \equiv \rho_{\text{S}}(\mathbf{r}) = \sum_{i=1}^{N} \left | \varphi_{i} \left ( \mathbf{x} \right ) \right |^{2} = \rho_{0}(\mathbf{r})
\nonumber \\
& %
\tikzmark{4}
V_{ne}[\rho ] =  \int -  \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}}  \sum_{i=1}^{N} \left | \varphi_{i} \left ( \mathbf{x}_{1} \right ) \right |^{2}  \mathrm{d}\mathbf {r}_{1}
 \nonumber \\
\tikzmark{5}
\Aboxed{
&V_{ne}[\rho ] =  - \sum_{i=1}^{N} \int   \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}}  \left | \varphi_{i} \left ( \mathbf{x}_{1} \right ) \right |^{2}  \mathrm{d}\mathbf {r}_{1}
}
 \nonumber \\
\tikzmark{6}
\Aboxed{
&V_{ee}[\rho ] = J[\rho ] + V_{\text{non-classical}}[\rho ]
}
 \nonumber \\
& \qquad \qquad J[\rho]  = \frac{1}{2} \int \int \frac{\rho(\mathbf{r}_{1})\rho(\mathbf{r}_{2})}{r_{12}} \mathrm{d}\mathbf {r}_{1} \mathrm{d}\mathbf {r}_{2}  \nonumber \\
& \qquad \qquad \qquad  \rho (\mathbf{r}) \equiv \rho_{\text{S}}(\mathbf{r}) = \sum_{i=1}^{N} \left | \varphi_{i} \left ( \mathbf{x} \right ) \right |^{2} = \rho_{0}(\mathbf{r})  \nonumber \\
&
\Aboxed{
J[\rho]  = \frac{1}{2} \sum_{i=1}^{N} \sum_{j=1}^{N} \int \int \left | \varphi_{i} (\mathbf{r}_{1}) \right |^{2} \frac{1}{r_{12}} \left | \varphi_{j} (\mathbf{r}_{2}) \right |^{2} \mathrm{d}\mathbf {r}_{1}  \mathrm{d}\mathbf {r}_{2}  }
\nonumber \\
\tikzmark{7}
\Aboxed{
&T[\rho ]  = T_{\text{S}}[\rho ] + T_{\text{C}}[\rho ]
}
 \nonumber \\
& \qquad \qquad T_{\text{S}}[\rho ] = -\frac{1}{2}\sum_{i=1}^{N} \expval{\nabla^{2}}{\varphi _{i}}
 \nonumber \\
\tikzmark{8}
\Aboxed{
E_{\text{XC}} [\rho] &= \left ( T[\rho] - T_{\text{S}}[\rho] \right ) + \left ( E_{ee}[\rho] - J[\rho] \right ) = T_{\text{C}}[\rho] + V_{\text{non-classsical}}[\rho]
}
\nonumber \\
\tikzmark{9}
\Aboxed{
E [\rho ] = & T_{\text{S}}[\rho ] + J[\rho] +  V_{ne}[\rho ] + E_{\text{XC}} [\rho]
}
\nonumber \\
%\begin{empheq}[box=\fbox]{align}
%\end{subequations}
%
\tikzmark{10}
E [\rho ] = &  -\frac{1}{2}\sum_{i=1}^{N} \expval{\nabla^{2}}{\varphi _{i}} +  \frac{1}{2} \sum_{i=1}^{N} \sum_{j=1}^{N} \int \int \left | \varphi_{i} (\mathbf{r}_{1}) \right |^{2} \frac{1}{r_{12}} \left | \varphi_{j} (\mathbf{r}_{2}) \right |^{2} \mathrm{d}\mathbf {r}_{1}  \mathrm{d}\mathbf {r}_{2}
\nonumber \\
&  + E_{\text{XC}} [\rho] - \sum_{i=1}^{N} \int   \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}}  \left | \varphi_{i} \left ( \mathbf{x}_{1} \right ) \right |^{2}  \mathrm{d}\mathbf {r}_{1}\\
%}
%\end{empheq}
\end{empheq}
\end{subequations}

\begin{tikzpicture}[remember picture, overlay]
\foreach \i [count=\ni] in {1,9,10}
    \node[draw, circle, inner sep=2pt, left=2mm of pic cs:\i, yshift=.5ex] (c\ni) {\ni};

\foreach \i in {2,...,4}
    \draw[->] ([shift={(2mm,-4.pt)}]pic cs:1) |- ([shift={(-4pt,.5ex)}]pic cs:\i);
\foreach \i in {5,...,8}
    \draw[->] ([shift={(2mm,-4.pt)}]pic cs:1) |- ([yshift=.5ex]pic cs:\i);
\end{tikzpicture}
1
  • 1. Circled numbers in next page because your equations doesn't fit in one page. Reduce top and bottom margins. 2. \Aboxed needs one and only one & inside. Change \Aboxed by \fbox{$\displaymath.... 3. Same problem with \Aboxed, it's only valid for one line equations. I don't know if a nested boxed empheq will work. 4. Use \captionof command to caption non floating environments.
    – Ignasi
    Oct 3, 2017 at 10:27

2 Answers 2

2

A simple solution with pstricks. I define an \Acolorboxed command, mimicked on the \Aboxed command from mathtools:

\documentclass[12pt]{article}
\usepackage{empheq}
\usepackage[margin=2.5cm]{geometry}
\usepackage[svgnames]{xcolor}
\usepackage{pst-node, multido}
\usepackage{auto-pst-pdf} % to compile with pdflatex
\makeatletter
\colorlet{framecolor}{Tomato}
\colorlet{bgcolor}{white}
\newcommand{\fcolorboxed}[3]{\fcolorbox{#1}{#2}{\m@th$\displaystyle#3$}}
\newcommand\Acolorboxed[1]{\let\bgroup{\romannumeral-`}\@Acolorboxed#1&&\ENDDNE}
\def\@Acolorboxed#1&#2&#3\ENDDNE{%
  \ifnum0=`{}\fi \setbox \z@
    \hbox{$\displaystyle#1{}\m@th$\kern\fboxsep \kern\fboxrule }%
    \edef\@tempa {\kern \wd\z@ &\kern -\the\wd\z@ \fboxsep
        \the\fboxsep \fboxrule \the\fboxrule }\@tempa \fcolorboxed{framecolor}{bgcolor}{#1#2}%
}
\makeatother
\def\ma-ht{\fontdimen22\textfont2}
\newcommand{\myeqlabel}[1]{\cput[linecolor=Tomato](-2em,0.7ex){\color{Tomato}#1}}
\begin{document}

\newcommand*\widefbox[1]{\hspace{2em}#1\hspace{2em}}
\begin{postscript}
\begin{subequations}
\begin{empheq}[box=\widefbox]{align}
\myeqlabel{1}\Acolorboxed{\rnode[b]{I}{a} = & b + c + d}
 \nonumber \\
\pnode[0,\ma-ht]{E1}& d = \begin{aligned}[t]
& f + g \\
& f = m/4
 \end{aligned} \nonumber\\
\pnode[0,\ma-ht]{E2}& d = \begin{aligned}[t]
 & j + kr \\
& j = n\cdot3/4
 \end{aligned} \nonumber \\
\pnode[0,\ma-ht]{E3} & d = l + o
 \nonumber \\
\pnode[0,\ma-ht]{E4}\Aboxed{& d = m/4 + q + n\cdot 3/4 + k + l + 0 }
 \nonumber \\
\pnode[0,\ma-ht]{E5}\Aboxed{&c = p + q}
 \nonumber \\
& \phantom{c ={}}
p = h/2 + h/2
 \nonumber \\
\pnode[0,\ma-ht]{E6}\Aboxed{&b = r + s}
 \nonumber \\
\pnode[0,\ma-ht]{E7}\Aboxed{&t = u + w}
 \nonumber \\[\baselineskip]
\myeqlabel{2}\Acolorboxed{a = & b + c + d}
 \nonumber \\[\baselineskip]
\myeqlabel{3}\Acolorboxed{a = & zz + z'}
 \nonumber \\ \nonumber\\
\end{empheq}
\end{subequations}
\psset{linewidth =0.4pt, linejoin=1, arrowinset=0.12, angleA=-90, angleB =-180, arrows=->, nodesepB=0.4em, nodesepA=1.44\fboxsep}
\multido{\i =1 + 1}{7}{\ncangle{I}{E\i}}
\end{postscript}

\end{document} 

enter image description here

Edit: A code for the real situation.

Note double integrals are not obtained with two \int commands (which results in a very bad spacing), but with \iint. We have a tighter spacing loading the esint package.

I also simplified the code for the maths part defining \dd for the differential symbol in integrals (with a better spacing), an \abs command for the absolute value and had to define \expval (which is not a standard LaTeX command) with the \DeclarePairedDelimiterX command from mathtools. Final remark: you don't have to load amsmath when you load empheq since it loads mathtools, which loads the former.

\documentclass[11pt]{article}
\usepackage{empheq, nccmath}
\usepackage[margin=2.5cm]{geometry}
\usepackage{caption}
\usepackage[svgnames]{xcolor}
\usepackage{pst-node, multido}
\usepackage{auto-pst-pdf} % to compile with pdflatex
\makeatletter
\colorlet{framecolor}{Tomato}
\colorlet{bgcolor}{white}
\newcommand{\fcolorboxed}[3]{\fcolorbox{#1}{#2}{\m@th$\displaystyle#3$}}
\newcommand\Acolorboxed[1]{\setlength{\fboxrule}{0.8pt}\let\bgroup{\romannumeral-`}\@Acolorboxed#1&&\ENDDNE}
\def\@Acolorboxed#1&#2&#3\ENDDNE{%
\ifnum0=`{}\fi \setbox \z@
\hbox{$\displaystyle#1{}\m@th$\kern\fboxsep \kern\fboxrule }%
\edef\@tempa {\kern \wd\z@ &\kern -\the\wd\z@ \fboxsep
    \the\fboxsep \fboxrule \the\fboxrule }\@tempa \fcolorboxed{framecolor}{bgcolor}{#1#2}%
}
\makeatother
\def\ma-ht{\fontdimen22\textfont2}
\newcommand{\myeqlabel}[1]{\cput[linecolor=Tomato](-1.5em,0.7ex){\color{Tomato}#1}}
\newcommand*{\dd}{\mathop{}\!\mathrm{d}} %
\usepackage{esint}
\DeclarePairedDelimiter\abs{\lvert}{\rvert}
\DeclarePairedDelimiterX\expval[2]{\langle}{\rangle}%
{#1\,\delimsize\vert\,\mathopen{}#2\,\delimsize\vert\,\mathopen{}#1}
\usepackage{tikz}
\usetikzlibrary{tikzmark, positioning}

\begin{document}

\newcommand*\widefbox[1]{\fbox{\hspace{2em}#1\hspace{2em}}}
\begin{postscript}
  \begin{subequations}
    \begin{empheq}[box=\widefbox]{align}
      %
      \myeqlabel{1}\Acolorboxed{\pnode[1em, -4.27ex]{I}%
        E [\rho ]= & \underbrace{T[\rho ] + V_{ee}[\rho ] }_{=\,\, F[\rho ]} + V_{ne}[\rho]
      }
      \nonumber \\
      & \begin{alignedat}{2}
        \pnode[0,\ma-ht]{E1}V_{ne}[\rho ] & = & & \int \rho \left ( \mathbf{r} \right ) v_{\text{ext}} \left ( \mathbf{r} \right ) \dd\mathbf {r} \\
        & & &\,v_{\text{ext}}\left ( \mathbf{r}_{i} \right ) = - \sum_{A=1}^{M} \frac{Z_{A}}{r_{iA}} \\[-1ex]
        \pnode[0,\ma-ht]{E2}
        V_{ne}[\rho ] & = & & \int - \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}} \rho \left ( \mathbf{r}_{1} \right ) \dd\mathbf {r}_{1} \\
        & & & \, \rho (\mathbf{r}) \equiv \rho_{\text{S}}(\mathbf{r}) = \sum_{i=1}^{N}\abs{\varphi_{i} \left ( \mathbf{x} \right )}^{2} = \rho_{0}(\mathbf{r}) \\
        \pnode[0,\ma-ht]{E3}
        V_{ne}[\rho ] & = & & \int - \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}} \sum_{i=1}^{N} \abs{\varphi_{i} \left (\mathbf{x}_{1} \right) }^{2} \mathrm{d}\mathbf {r}_{1}
      \end{alignedat}\nonumber \\
      \pnode[0,\ma-ht]{E4} \Aboxed{
      &V_{ne}[\rho ]= - \sum_{i=1}^{N} \int \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}} \abs{\varphi_{i} \left ( \mathbf{x}_{1} \right )}^{2} \dd\mathbf {r}_{1}
      }
      \nonumber\\
      \pnode[0,\ma-ht]{E5}
      \Aboxed{%
        & V_{ee}[\rho ] = J[\rho ] + V_{\text{non-classical}}[\rho ]
      } \nonumber \\
      & \begin{aligned}\phantom{V_{ee}[\rho ] ={}} J[\rho] = {} & \mfrac{1}{2} \iint \frac{\rho (\mathbf{r}_{1})\rho (\mathbf{r}_{2})}{r_{12}} \dd\mathbf {r}_{1} \dd\mathbf {r}_{2} \\
                                                 & \rho (\mathbf{r}) \equiv \rho_{\text{S}}(\mathbf{r}) = \sum_{i=1}^{N} \abs{\varphi_{i} \left ( \mathbf{x} \right )}^{2} = \rho_{0}(\mathbf{r}) r \\
        \Aboxed{
        J[\rho]= & \mfrac{1}{2} \sum_{i=1}^{N} \sum_{j=1}^{N} \iint \abs*{\varphi_{i} (\mathbf{r}_{1})}^{2} \frac{1}{r_{12}} \abs*{\varphi_{j} (\mathbf{r}_{2})}^{2} \dd\mathbf {r}_{1} \dd\mathbf {r}_{2}}
      \end{aligned}
      \nonumber \\
      \pnode[0,\ma-ht]{E6}
      \Aboxed{
      &T[\rho ] = T_{\text{S}}[\rho ] + T_{\text{C}}[\rho ]
      }
      \nonumber \\
      %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
      &\phantom{T[\rho ] ={}} T_{\text{S}}[\rho ] = -\mfrac{1}{2}\sum_{i=1}^{N}\expval*{\nabla^{2}}{\varphi _{i}}%
      \nonumber \\
      \pnode[0,\ma-ht]{E7}
      \Aboxed{
        & E_{\text{XC}} [\rho] = \left ( T[\rho] - T_{\text{S}}[\rho] \right ) + \left ( E_{ee}[\rho] - J[\rho] \right ) = T_{\text{C}}[\rho] + V_{\text{non-classical}}[\rho]
      }
      \nonumber \\
      \myeqlabel{2}
      \Acolorboxed{
        E [\rho ] &= T_{\text{S}}[\rho ] + J[\rho] + V_{ne}[\rho ] + E_{\text{XC}} [\rho]
      }
      \nonumber \\
      \myeqlabel{3}
      \Acolorboxed{E [\rho ] &=\begin{aligned}[t]-\mfrac{1}{2}\sum_{i=1}^{N} %
          \expval*{\nabla^{2}}{\varphi _{i}} + \mfrac{1}{2} \sum_{i=1}^{N} \sum_{j=1}^{N} \iint \abs{\varphi_{i} (\mathbf{r}_{1})}^{2} \frac{1}{r_{12}} \abs*{\varphi_{j} (\mathbf{r}_{2})}^{2} \dd\mathbf {r}_{1} \dd\mathbf {r}_{2}
          \\[-1.5ex]
          + E_{\text{XC}} [\rho] - \sum_{i=1}^{N} \int \sum_{A=1}^{M} \frac{Z_{A}}{r_{1A}} \abs{\varphi_{i} \left ( \mathbf{x}_{1} \right )}^{2} \dd\mathbf {r}_{1}
        \end{aligned}\nonumber} \\
    \end{empheq}
  \end{subequations}
  %
  \psset{linewidth=0.6pt, linejoin=1, linecolor=LightSteelBlue, linearc=0.08, arrowinset=0, angleA=-90, angleB=-180, arrows=->, nodesepB=0.4em, nodesepA=1.44\fboxsep}
  \multido{\i=1+1}{3}{\ncangle{I}{E\i}}
  \psset{nodesepB=0em}
  \multido{\i=4+1}{4}{\ncangle{I}{E\i}}
  \captionof{figure}{Some nice equations}
\end{postscript}

\end{document} 

enter image description here

4
  • Thanks a lot for your approach. I am trying to apply this to the real example (see below), but I cannot manage to obtain the result. For the moment, we can forget about the coloured scheme, it is more important to get the scheme right. I would appreciate if you could have a look. Million Thanks
    – DavidC.
    Oct 2, 2017 at 21:14
  • Did you compile with pdflatex -shell-escape?
    – Bernard
    Oct 2, 2017 at 21:20
  • Yes, exactly, I compiled with pdflatex -shell-escape
    – DavidC.
    Oct 2, 2017 at 21:29
  • 1
    @DavidC.: I've added a code for your real situation. Please check if it compiles for you.
    – Bernard
    Oct 3, 2017 at 13:09
3

You can use tikzmarks inside the empheq environments and later on, draw the numbers and connections:

\documentclass[12pt]{article}    
\usepackage{empheq}
\usepackage[left=2.5cm,top=2.5cm,right=2.5cm,bottom=2.5cm]{geometry}
\usepackage{amsmath}
\usepackage{tikz}
\usetikzlibrary{tikzmark, positioning}

\begin{document}

\newcommand*\widefbox[1]{\fbox{\hspace{2em}#1\hspace{2em}}}

\begin{subequations}
\begin{empheq}[box=\widefbox]{align}
%
\tikzmark{1}\Aboxed{a = & b + c + d}
 \nonumber \\
& \tikzmark{2} d = f + g
 \nonumber \\
& \qquad
f = m/4
 \nonumber \\
&
\tikzmark{3} d = j + k
 \nonumber \\
& \qquad
j = n\cdot3/4
 \nonumber \\
&\tikzmark{4} d = l + o
 \nonumber \\
\tikzmark{5}\Aboxed{& d = m/4 + q + n\cdot 3/4 + k + l + 0 }
 \nonumber \\
\tikzmark{6}\Aboxed{&c = p + q}
 \nonumber \\
& \qquad
p = h/2 + h/2
 \nonumber \\
\tikzmark{7}\Aboxed{&b = r + s}
 \nonumber \\
\tikzmark{8}\Aboxed{&t = u + w}
 \nonumber \\  \nonumber\\
\tikzmark{9}\Aboxed{a = & b + c + d}
 \nonumber \\ \nonumber\\
\tikzmark{10}\Aboxed{a = & zz + z'}
 \nonumber \\ \nonumber
\end{empheq}
\end{subequations}

\begin{tikzpicture}[remember picture, overlay]
\foreach \i [count=\ni] in {1,9,10}
    \node[draw, circle, inner sep=2pt, left=2mm of pic cs:\i, yshift=.5ex] (c\ni) {\ni};

\foreach \i in {2,...,4}
    \draw[->] ([shift={(2mm,-4.pt)}]pic cs:1) |- ([shift={(-4pt,.5ex)}]pic cs:\i);
\foreach \i in {5,...,8}
    \draw[->] ([shift={(2mm,-4.pt)}]pic cs:1) |- ([yshift=.5ex]pic cs:\i);
\end{tikzpicture}

\end{document}

enter image description here

Update: red color addition.

With Bernard's Acolorboxed definition and a little change in round labels, it's easy to get the desired result:

enter image description here

\documentclass[12pt]{article}    
\usepackage{empheq}
\usepackage[left=2.5cm,top=2.5cm,right=2.5cm,bottom=2.5cm]{geometry}
\usepackage{amsmath}
\usepackage{tikz}
\usetikzlibrary{tikzmark, positioning}

\makeatletter
\colorlet{framecolor}{red}
\colorlet{bgcolor}{white}
\newcommand{\fcolorboxed}[3]{\fcolorbox{#1}{#2}{\m@th$\displaystyle#3$}}
\newcommand\Acolorboxed[1]{\let\bgroup{\romannumeral-`}\@Acolorboxed#1&&\ENDDNE}
\def\@Acolorboxed#1&#2&#3\ENDDNE{%
  \ifnum0=`{}\fi \setbox \z@
    \hbox{$\displaystyle#1{}\m@th$\kern\fboxsep \kern\fboxrule }%
    \edef\@tempa {\kern \wd\z@ &\kern -\the\wd\z@ \fboxsep
        \the\fboxsep \fboxrule \the\fboxrule }\@tempa \fcolorboxed{framecolor}{bgcolor}{#1#2}%
}
\makeatother

\begin{document}

\newcommand*\widefbox[1]{\fbox{\hspace{2em}#1\hspace{2em}}}

\begin{subequations}
\begin{empheq}[box=\widefbox]{align}
%
\tikzmark{1}\Acolorboxed{a = & b + c + d}
 \nonumber \\
& \tikzmark{2} d = f + g
 \nonumber \\
& \qquad
f = m/4
 \nonumber \\
&
\tikzmark{3} d = j + k
 \nonumber \\
& \qquad
j = n\cdot3/4
 \nonumber \\
&\tikzmark{4} d = l + o
 \nonumber \\
\tikzmark{5}\Aboxed{& d = m/4 + q + n\cdot 3/4 + k + l + 0 }
 \nonumber \\
\tikzmark{6}\Aboxed{&c = p + q}
 \nonumber \\
& \qquad
p = h/2 + h/2
 \nonumber \\
\tikzmark{7}\Aboxed{&b = r + s}
 \nonumber \\
\tikzmark{8}\Aboxed{&t = u + w}
 \nonumber \\  \nonumber\\
\tikzmark{9}\Acolorboxed{a = & b + c + d}
 \nonumber \\ \nonumber\\
\tikzmark{10}\Acolorboxed{a = & zz + z'}
 \nonumber \\ \nonumber
\end{empheq}
\end{subequations}

\begin{tikzpicture}[remember picture, overlay]
\foreach \i [count=\ni] in {1,9,10}
    \node[draw, red, circle, inner sep=2pt, left=2mm of pic cs:\i, yshift=.5ex] (c\ni) {\ni};

\foreach \i in {2,...,4}
    \draw[->] ([shift={(2mm,-4.pt)}]pic cs:1) |- ([shift={(-4pt,.5ex)}]pic cs:\i);
\foreach \i in {5,...,8}
    \draw[->] ([shift={(2mm,-4.pt)}]pic cs:1) |- ([yshift=.5ex]pic cs:\i);
\end{tikzpicture}

\end{document}
3
  • Thanks a lot for all your effort. I have been quite a long time trying to apply this to the real example, but I have encountered the difficulties explained below. Could you please help me to achieve this? Again, thank you very much
    – DavidC.
    Oct 2, 2017 at 21:10
  • 1
    @DavidC. Answer updated with red labels and boxes.
    – Ignasi
    Oct 3, 2017 at 7:32
  • Thank you for your update. Please see updated post: I am finding difficulties to apply this code to the real example. Thanks again
    – DavidC.
    Oct 3, 2017 at 9:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .