相关概念
Error Reduction (one-sided case)
基本概念
$\forall x\in \left\{ 0,1 \right\} ^*:f\left( x \right) =1\Rightarrow \mathscr{A}\left( x \right) =1$
$\forall x\in \left\{ 0,1 \right\} ^*:f\left( x \right) =0\Rightarrow Pr\left( \mathscr{A}^n\left( x \right) =0 \right) \ge p$
定义一个决定性问题:
$$
f:\left\{ 0,1 \right\} ^*=\left\{ 0,1 \right\}
$$
有一个具有单边错误的蒙特卡洛随机算法$\mathscr{A}$,算法满足:
算法$ \mathscr{A}^n$过程为,相互独立地执行$\mathscr{A}$共$n$次,取所有结果的且$\left( \land \right) $进行输出。
进而可以推得算法犯错的概率$P$满足:
$$
P\le Pr\left( \mathscr{A}^n\left( x \right) =1 \right) \le \left( 1-p \right) ^n
$$
故执行$n$次算法以后失败的概率将收敛于任意正数$\varepsilon $,其中$n$的取值为:
$$
\left( 1-p \right) ^n=\varepsilon \Rightarrow e^{-np}\approx \varepsilon \Rightarrow n\approx \frac{1}{p}\ln \left( \frac{1}{\varepsilon} \right)
$$