A Monte Carlo Approach to Options

最新推荐文章于 2023-04-04 19:41:13 发布

Mochaaaa

最新推荐文章于 2023-04-04 19:41:13 发布

阅读量342

点赞数

分类专栏：统计基础

本文链接：https://blog.csdn.net/Mochaaaa/article/details/53292978

版权

统计基础专栏收录该内容

2 篇文章 0 订阅

订阅专栏

Basic

Let $Y$ denote a continuous random variable. Using Monte Carlo method, we first consider the integral

g ̂ = 1 n \sum g (y i)

$\begin{equation} \hat g=\frac{1}{n}\sum g(y_{i}) \end{equation}$
with variance

s ̂ 2 = 1 n - 1 \sum i - 1 n (g (y i) - g ̂) 2

$\begin{equation} \hat s^{2}=\frac{1}{n-1}\sum^{n}_{i-1}\bigg(g(y_{i})-\hat g\bigg)^{2} \end{equation}$
For large sample size,

n−1 $n-1$ in equation above is replaced by n. By the Central Limit Theorem, we have

g ̂ - g ¯ s ̂ 2 n ‾ ‾ ‾ \sqrt \sim N (0, 1)

$\begin{equation} \frac{\hat g - \bar g}{\sqrt{\frac{\hat s^{2}}{n}}} \sim N(0,1) \end{equation}$
Therefore, we know that in order to reduce the standard error of

ĝ $\hat g$ ,which equals

ŝ 2n‾‾‾√ $\sqrt{\frac{\hat s^{2}}{n}}$ , we need either increase the number of trials

n $n$ or reduce the size of

ŝ $\hat s$ .

Variance Reduction Techniques: Control Variate method

For a given n,
generate $y_{i}$ from $g(y)$
generate $Y_{i}$ from $h(y)$
$\hat g$ = $\int g(y)f(y)dy$
$\hat G =\int g(y)h(y)dy$
$g^{*}=G +(\hat g - \hat G)$ , which is a revised estimate of $\bar g$

Variance Reduction Techniques: Antithetic method

Generate path $y_{1},y_{2},...,y_{n}$ through sequence $u_{1},u_{2},...,u_{n}$ from standard uniform distribution
Generate another path through $(1-u_{1}),(1-u_{2}),...,(1-u_{n})$
Suppose the estimate of (1) using the first path is $g(u)$ and the estimate of (1) using the second path is (1-u)
Then the combined estimator $g'=\frac{1}{2}[g(u)+g(1-u)]$