我遇到了这个图的标题的一个恼人的问题：
\documentclass[11pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{fix-cm,relsize}
\usepackage[sort&compress,numbers]{natbib}
\usepackage{amsmath,kbordermatrix,amsthm,amssymb,subfigure,mathrsfs,rotating,array,graphicx,url,booktabs, multirow,eurosym,color}
\usepackage[shortlabels]{enumitem}
\usepackage{boldline}
\usepackage[thinlines]{easytable}
\usepackage{makecell}
\setdescription{style=multiline,topsep=10pt,leftmargin=2.1cm,font=\normalfont}
\usepackage[font=footnotesize,labelfont=bf,singlelinecheck=false]{caption}
\usepackage[a4paper, includeheadfoot, left=2cm,right=2cm,top=1cm,bottom=1cm,headheight=0cm]{geometry}
\usepackage[pdftex,pdfstartview=FitBH,bookmarks=true,bookmarksnumbered=true]{hyperref}
\usepackage{pdflscape}
\begin{document}
\renewcommand{\kbldelim}{(}% Left delimiter
\renewcommand{\kbrdelim}{)}% Right delimiter
\begin{figure}[!b]
\centering
\parbox{5cm}{
\begin{eqnarray*}
\kbordermatrix{
& t_1 & t_2 & t_3 & t_4 & t_5 \\
t_1 & 0 & 2 & 3 & 3 & 2 \\
t_2 & 2 & 0 & 2 & 2 & 3 \\
t_3 & 3 & 2 & 0 & 2 & 3 \\
t_4 & 3 & 2 & 2 & 0 & 2 \\
t_5 & 2 & 3 & 3 & 2 & 0
}
\end{eqnarray*}
}
\quad
\begin{minipage}{5cm}
\centering
\begin{tabular}{|c|ccc|}
\hline
Taxon & \multicolumn{3}{c|}{\small{Sequence}} \\
\hline
$t_1$ & A & A & A \\%& \ldots\\
$t_2$ & A & C & C \\%& \ldots\\
$t_3$ & C & G & C \\%& \ldots\\
$t_4$ & C & C & G \\%& \ldots\\
$t_5$ & G & A & G \\%& \ldots\\
\hline
\end{tabular}
\end{minipage}
\caption{The Balanced Minimum Evolution Model (BMEM) is an alternative example of a phylogenetic estimation model that can be described as an optimization problem. The BMEM arises from the distance methods and, as for the MPM, it represents a phylogeny of a given set $\Gamma$ of taxa as an UBT having $\Gamma$ as leaf-set. The BMEM assumes the existence of a measure of dissimilarity (or \emph{distance}) $d_{ij}$ for each pair of distinct taxa $i,j\in\Gamma$. When the observed heritable traits are molecular sequences, such a measure can be computed by using, e.g., one of the nucleotide substitution models described in \citep{Daniele1, FelsenBook, HKY, Lanave, JC, Kimura, Rodri, Waddell, Beyer,Kidd,Fitch03, Galtier01, Huelsen05, Lopez,Kidd}. As an example, if we would just use the Hamming distance to measure the dissimilarity among the molecular sequences shown in Figure~\protect\ref{Figure02}, we would obtain the symmetric distance matrix shown on the left of this figure. Based on this information, the BMEM states that the phylogeny of a given set $\Gamma$ of taxa can be computed by solving the following combinatorial optimization problem:
\[
\protect\min_{T\in \mathcal{T} } \; \displaystyle \sum_{\substack{i,j\in\Gamma\\i\neq j}} \frac{d_{ij}}{2^{\tau_{ij}}}
\]
where $\mathcal{T}$ represents the set of the $(2n-5)!!$ possible phylogenies of $\Gamma$ \citep{FelsenBook,Daniele10} and the \emph{path-length} (or \emph{topological distance}) $\tau_{ij}$ represents the number of edges belonging to the (unique) path from taxon $i$ to taxon $j$ in a phylogeny $T\in\mathcal{T}$ \citep{Pauplin,GascuelBook,Daniele11}. The optimal solution to this optimization problem when considering the above distance matrix is shown on the right of this figure.}\label{Figure03}
\end{figure}
\end{document}
具体来说，latex 写道，存在以下错误：
/Users/darkstar/Dropbox/Ricerca/Progetti/2017/2017 - 01 - ERC/A - Latex sources/ERCB2.tex:125: Argument of \caption@ydblarg has an extra }. [...ix is shown on the right of this figure.}] /Users/darkstar/Dropbox/Ricerca/Progetti/2017/2017 - 01 - ERC/A - Latex sources/ERCB2.tex:125: Paragraph ended before \caption@ydblarg was complete. [...ix is shown on the right of this figure.}] /Users/darkstar/Dropbox/Ricerca/Progetti/2017/2017 - 01 - ERC/A - Latex sources/ERCB2.tex:125: Missing $ inserted. [...ix is shown on the right of this figure.}] /Users/darkstar/Dropbox/Ricerca/Progetti/2017/2017 - 01 - ERC/A - Latex sources/ERCB2.tex:125: Missing $ inserted. [...ix is shown on the right of this figure.}]
但是，所有 $ 和 } 都是正确的！问题出在哪里？
答案1
请始终发布重现错误的完整文档，但在标准类中\caption设置为水平模式来测量其宽度，这将导致显示数学收缩等\[失败。
caption禁用该功能的最简单的方法是使用
singlelinecheck=false
运行没有错误
\documentclass[11pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{fix-cm,relsize}
\usepackage[sort&compress,numbers]{natbib}
\usepackage{amsmath,kbordermatrix,amsthm,amssymb,subfigure,mathrsfs,rotating,array,graphicx,url,booktabs, multirow,eurosym,color}
\usepackage[shortlabels]{enumitem}
\usepackage{boldline}
\usepackage[thinlines]{easytable}
\usepackage{makecell}
\setdescription{style=multiline,topsep=10pt,leftmargin=2.1cm,font=\normalfont}
\usepackage[font=footnotesize,labelfont=bf,singlelinecheck=false]{caption}
\usepackage[a4paper, includeheadfoot, left=2cm,right=2cm,top=1cm,bottom=1cm,headheight=0cm]{geometry}
\usepackage[pdftex,pdfstartview=FitBH,bookmarks=true,bookmarksnumbered=true]{hyperref}
\usepackage{pdflscape}
\begin{document}
\renewcommand{\kbldelim}{(}% Left delimiter
\renewcommand{\kbrdelim}{)}% Right delimiter
\begin{figure}[!b]
\centering
\parbox{5cm}{
\[
\kbordermatrix{
& t_1 & t_2 & t_3 & t_4 & t_5 \\
t_1 & 0 & 2 & 3 & 3 & 2 \\
t_2 & 2 & 0 & 2 & 2 & 3 \\
t_3 & 3 & 2 & 0 & 2 & 3 \\
t_4 & 3 & 2 & 2 & 0 & 2 \\
t_5 & 2 & 3 & 3 & 2 & 0
}
\]
}
\quad
\begin{minipage}{5cm}
\centering
\begin{tabular}{|c|ccc|}
\hline
Taxon & \multicolumn{3}{c|}{\small{Sequence}} \\
\hline
$t_1$ & A & A & A \\%& \ldots\\
$t_2$ & A & C & C \\%& \ldots\\
$t_3$ & C & G & C \\%& \ldots\\
$t_4$ & C & C & G \\%& \ldots\\
$t_5$ & G & A & G \\%& \ldots\\
\hline
\end{tabular}
\end{minipage}
\caption[The Balanced Minimum Evolution Model]{The Balanced Minimum Evolution Model (BMEM) is an alternative example of a phylogenetic estimation model that can be described as an optimization problem. The BMEM arises from the distance methods and, as for the MPM, it represents a phylogeny of a given set $\Gamma$ of taxa as an UBT having $\Gamma$ as leaf-set. The BMEM assumes the existence of a measure of dissimilarity (or \emph{distance}) $d_{ij}$ for each pair of distinct taxa $i,j\in\Gamma$. When the observed heritable traits are molecular sequences, such a measure can be computed by using, e.g., one of the nucleotide substitution models described in \citep{Daniele1, FelsenBook, HKY, Lanave, JC, Kimura, Rodri, Waddell, Beyer,Kidd,Fitch03, Galtier01, Huelsen05, Lopez,Kidd}. As an example, if we would just use the Hamming distance to measure the dissimilarity among the molecular sequences shown in Figure~\protect\ref{Figure02}, we would obtain the symmetric distance matrix shown on the left of this figure. Based on this information, the BMEM states that the phylogeny of a given set $\Gamma$ of taxa can be computed by solving the following combinatorial optimization problem:
\[
\min_{T\in \mathcal{T} } \; \sum_{\substack{i,j\in\Gamma\\i\neq j}} \frac{d_{ij}}{2^{\tau_{ij}}}
\]
where $\mathcal{T}$ represents the set of the $(2n-5)!!$ possible phylogenies of $\Gamma$ \citep{FelsenBook,Daniele10} and the \emph{path-length} (or \emph{topological distance}) $\tau_{ij}$ represents the number of edges belonging to the (unique) path from taxon $i$ to taxon $j$ in a phylogeny $T\in\mathcal{T}$ \citep{Pauplin,GascuelBook,Daniele11}. The optimal solution to this optimization problem when considering the above distance matrix is shown on the right of this figure.}\label{Figure03}
\end{figure}
\end{document}
答案2
我认为在 的参数中包含多行显示（包括显示方程式）没有任何理由\caption。简短的标题，例如，
\caption{Balanced Minimum Evolution Model (BMEM)}
对读者来说几乎更有帮助。其余材料可以排版为普通图例。以这种方式简化标题还可以摆脱这两个\protect语句，并自动摆脱您遇到的错误消息。
另外，不要使用eqnarray*。一个简单的\[ ...\]构造就足够了。
\documentclass[demo]{article} % remove 'demo' option in real doc.
\usepackage{graphicx,natbib,amsmath,kbordermatrix}
\begin{document}
\begin{figure}[!b]
\begin{minipage}{5cm}
\[
\kbordermatrix{
& t_1 & t_2 & t_3 & t_4 & t_5 \\
t_1 & 0 & 2 & 3 & 3 & 2 \\
t_2 & 2 & 0 & 2 & 2 & 3 \\
t_3 & 3 & 2 & 0 & 2 & 3 \\
t_4 & 3 & 2 & 2 & 0 & 2 \\
t_5 & 2 & 3 & 3 & 2 & 0
}
\]
\end{minipage}
\quad
\begin{minipage}{5cm}
\centering
\includegraphics[scale=0.5]{Fig03-SolutionBME}
\end{minipage}
\caption{Balanced Minimum Evolution Model (BMEM)} \label{Figure03}
\bigskip
The Balanced Minimum Evolution Model (BMEM) is an alternative example of a phylogenetic estimation model that can be described as an optimization problem. The BMEM arises from the distance methods and, as for the MPM, it represents a phylogeny of a given set~$\Gamma$ of taxa as an UBT having $\Gamma$ as leaf-set. The BMEM assumes the existence of a measure of dissimilarity (or \emph{distance}) $d_{ij}$ for each pair of distinct taxa~$i,j\in\Gamma$. When the observed heritable traits are molecular sequences, such a measure can be computed by using, e.g., one of the nucleotide substitution models described in \citep{Daniele1, FelsenBook, HKY, Lanave, JC, Kimura, Rodri, Waddell, Beyer, Kidd, Fitch03, Galtier01, Huelsen05, Lopez, Kidd}. As an example, if we would just use the Hamming distance to measure the dissimilarity among the molecular sequences shown in Figure~\ref{Figure02}, we would obtain the symmetric distance matrix shown on the left of this figure. Based on this information, the BMEM states that the phylogeny of a given set~$\Gamma$ of taxa can be computed by solving the following combinatorial optimization problem:
\[
\min_{T\in\mathcal{T}}
\sum_{\substack{i,j\in\Gamma \\ i\neq j}}
\frac{d_{ij}}{2^{\tau_{ij}}}
\]
where $\mathcal{T}$ represents the set of the $(2n-5)!!$ possible phylogenies of~$\Gamma$ \citep{FelsenBook,Daniele10} and the \emph{path-length} (or \emph{topological distance}) $\tau_{ij}$ represents the number of edges belonging to the (unique) path from taxon~$i$ to taxon~$j$ in a phylogeny $T\in\mathcal{T}$ \citep{Pauplin, GascuelBook, Daniele11}. The optimal solution to this optimization problem when considering the above distance matrix is shown on the right of this figure.
\end{figure}
\end{document}