大偏差最优路径下不同状态重要抽样-统计学专业论文.docx

I I 大偏差最优路径下的不同状态重要抽样 中文摘要 中文摘要 小概率事件的有效模拟是概率统计中的一个热点研究领域。小概率 事件出现在许多应用中，如电信、计算机、通信网络、保险、金融等。 小概率事件的模拟方法有许多，常用的有重要抽样、条件蒙特卡洛、交 叉熵、大偏差技术及分裂方法。 本文讨论在轻尾随机变量下一种有效的小概率事件模拟方法。该方 法主要依赖于大偏差技术和重要抽样方法的结合。首先根据大偏差理论 得出小概率事件发生的最优路径。其次根据大偏差最优路径与重要抽样 中指数测度变换方法的内在联系，得出相应指数测度变换。最后根据实 际情况选择采用不同状态下的重要抽样。本文给出大偏差最优路径下的 不同状态重要抽样方法的具体步骤，通过 SAS 软件模拟，应用于一个轻 尾和例子及一个相对复杂的数值障碍期权例子。 关键词：小概率事件；大偏差技术；重要抽样；最优路径；状态相依 作 者：蒋静文 指导老师：汪四水 II II 英文摘要 大偏差最优路径下的不同状态重要抽样 Important sampling of different states with optimal-path based on large-deviations Abstract Efficient simulation of rare events has been an active research area in applied probability. Rare events arise in many applications including telecommunications, computer and communication networks, insurance and finance. There are many different methods of rare events simulation, such as importance sampling, conditional Monte Carlo, cross-entropy, large-deviations approach and splitting. In this paper, we consider one of efficient simulations of rare events in light-tailed settings. It relies on the combination between large-deviations and importance sampling. Firstly, we get the optimal path by using the large-deviations approach. Secondly, owe to the interplay between large deviations techniques and importance sampling, we can calculate the exponential change of measure corresponding to the optimal path. Finally, we choose one of importance sampling algorithms about different states. In the paper, we set up two examples in SAS by using the method, including the one about the sum of light-tailed and a complex digital barrier options. Keywords: Rare events; Importance sampling; Large deviations; Optimal path; State-dependent. Written by Jiang Jingwen Supervised by Wang Sishui 目 录 HYPERLINK \l _bookmark0 第一章 引 言 1 HYPERLINK \l _bookmark1 §1.1 小概率事件 1 HYPERLINK \l _bookmark2 §1.2 有效性的概念 2 HYPERLINK \l _bookmark3 §1.3 重要抽样 3 HYPERLINK \l _bookmark4 §1.3.1 重要抽样基本概念 3 HYPERLINK \l _bookmark5 §1.3.2 指数测度变换 4 HYPERLINK \l _bookmark6 §1.4 相依状态重要抽样与独