Important Sampling Scheme Based on Minimal Cuts for Stochastic Network Unreliability Estimation
博士 === 國立臺灣科技大學 === 資訊管理系 === 101 === Computer simulation is often used to estimate the system unavailability of a stochastic network. However, sampling efforts required for estimating unavailability of a highly reliable network become formidable. An importance sampling scheme based on a subset of t...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2013
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/69031821584278262877 |
| Summary: | 博士 === 國立臺灣科技大學 === 資訊管理系 === 101 === Computer simulation is often used to estimate the system unavailability of a stochastic network. However, sampling efforts required for estimating unavailability of a highly reliable network become formidable. An importance sampling scheme based on a subset of the minimal cuts of the network is proposed. The idea is to keep the importance distribution close to the original state distribution over the region of system failure and directly generate the failure states. The proposed estimator is shown to be unbiased and the variance of the estimator is derived. Empirical results show that for highly reliable networks the proposed estimator can achieve substantial variance reduction.
|
|---|