\begin{enumerate}[fullwidth,itemindent=2em,label={\bf 步骤~ \arabic*.}]
\end{enumerate}

实现上面列表样式的代码如下：

\begin{enumerate}[fullwidth,itemindent=2em,label={\bf Step~ \arabic*.}]
\item Generating data.
  \begin{enumerate}[label={\bf (\alph*)},leftmargin=6em]
    \item Given $\bm\Phi,\bm\Sigma,\bm\mu_x,\bm\mu_y,n_1,n_2$. ($\bm\mu_x$ was set to $\bm 0$ in this paper)
    \item Sample randomly stochastic disturbance term $\{\bm\varepsilon_t,t=1,\cdots,n_1+n_2\}$ from a normal distribution with mean $\bm 0$ and covariance matrix $\bm \Sigma$.
    \item Set $\bm x_0=\bm 0$, then generate data of Phase I $\{\bm x_t,t=1,\cdots,n_1\}$ by equation \eqref{eq:chap5_16} with $\bm\Phi,\bm\mu_x,\bm\varepsilon_t,t=1,\cdots,n_1$.
    \item Set $\bm y_0=\bm x_{n_1}$, then generate data of Phase II $\{\bm y_t,t=n_1+1,\cdots,n_1+n_2\}$ by equation \eqref{eq:chap5_17} with $\bm\Phi,\bm\mu_y,\bm\varepsilon_t,t=n_1+1,\cdots,n_1+n_2$.
  \end{enumerate}
\item Diagnosing Process.
  \begin{enumerate}[label={\bf (\alph*)},leftmargin=6em]
    \item Fit VAR(1) model to data generated from Step 1, and then
    structure pseudo-residual sequences $\{\bm e_{t,x},t=1,\cdots,n_1\}$ and $\{\bm e_{t}^s,t=n_1+1,\cdots,n_1+n_2\}$ by the Section 2.2.
    \item Adopt three methods (BDM, LEB, Step-Down) to diagnose process shifts using raw data $\{\bm x_t,\bm y_t\}$ and pseudo-residual sequences $\{\bm e_{t,x},\bm e_{t}^s\}$, respectively.
    \item Make decision by rules of different diagnostic procedures.
  \end{enumerate}
\item Repeat Step 1 and Step 2 with $n$ times to calculate "C" and "ENE".
\end{enumerate}	

参数解释：