# How to write this math term? with cases it isn't working

I can't find a way to write this in Latex. Any hints where to start?

• Welcome to TeX.SE. – Mico Apr 18 at 5:22

A combination of a drcases environment (provided by the mathtools package) and a plain tabular environment gets the job done:

\documentclass{article}
\usepackage{mathtools} % for 'drcases' environment
\begin{document}

$\begin{drcases} g^l_{ij} = \frac{\partial e}{\partial w^l_{ij}} = \dots\\ w^l_{ij} = \dots \end{drcases} \qquad \begin{tabular}{@{}l} for l=\dots\\ for i=\dots\\ for j=\dots \end{tabular}$
\end{document}


A combination of aligned and alignedat so all equals signs are aligned.

\documentclass{article}
\usepackage{amsmath}

\newcommand{\pder}[2]{\frac{\partial #1}{\partial #2}}

\begin{document}

\left. \begin{aligned} g_{ij}^l &= \pder{e}{w_{ij}^l} = \mathcal{A}(I_i^l)\delta_i^{l+1} \\[2ex] w_{ij}^l &= w_{ij}^l-rg_{ij}^l \end{aligned} \right\} \qquad \begin{alignedat}{2} &\text{for} &\quad l &= 1,2,3,\dots, L,\\ &\text{for} & i &= 0,1,2,\dots, N^{l+1},\\ &\text{for} & j &= 1,2,3,\dots, N^{l+1}. \end{alignedat}
\end{document}


The bottom equality in the left part is suspicious, check it.

Similar to Mico's approach but with the equal sign aligned.

\documentclass{article}
\usepackage{mathtools}
\begin{document}
\begin{align*}
\left.\begin{aligned}
g_{ij}^i &=\frac{\partial e}{\partial w_{ij}^I}=A(I_i)\delta^{\ell+1}_j\\
w_{ij}^i &=w_{ij}^\ell-r\,g_{ij}^\ell\\
\text{for $\ell=1,2,3,\dots L$,}\\
\text{for $i=0,1,2,\dots N^{\ell+1}$,}\\
\text{for $j=1,2,3,\dots N^{\ell+1+2}$}

• +1. :-) I think, though, that it's supposed to be w^l (or w^\ell, if you prefer) rather than either w^I or w^i... – Mico Apr 18 at 6:16