Markov Chain Monte Carlo Simulations And Their Statistical Analysis: With Web-based Fortran Code

Bernd A. Berg

  • 出版商: World Scientific Pub
  • 出版日期: 2004-10-14
  • 售價: $1,400
  • 貴賓價: 9.8$1,372
  • 語言: 英文
  • 頁數: 378
  • 裝訂: Paperback
  • ISBN: 9812389350
  • ISBN-13: 9789812389350
  • 相關分類: 程式語言
  • 下單後立即進貨 (約5~7天)





This book teaches modern Markov chain Monte Carlo (MC) simulation techniques step by step. The material should be accessible to advanced undergraduate students and is suitable for a course. It ranges from elementary statistics concepts (the theory behind MC simulations), through conventional Metropolis and heat bath algorithms, autocorrelations and the analysis of the performance of MC algorithms, to advanced topics including the multicanonical approach, cluster algorithms and parallel computing. Therefore, it is also of interest to researchers in the field. The book relates the theory directly to Web-based computer code. This allows readers to get quickly started with their own simulations and to verify many numerical examples easily. The present code is in Fortran 77, for which compilers are freely available. The principles taught are important for users of other programming languages, like C or C++.


Table of Contents:

  • Sampling, Statistics and Computer Code
  • Error Analysis for Independent Random Variables
  • Markov Chain Monte Carlo
  • Error Analysis for Markov Chain Data
  • Advanced Monte Carlo
  • Parallel Computing
  • Conclusions, History and Outlook



這本書逐步教授現代馬可夫鏈蒙地卡羅(MC)模擬技術。內容應該對高年級本科生易於理解,並適合作為一門課程。從基本統計概念(MC模擬背後的理論)開始,通過常規的Metropolis和熱浴算法、自相關和MC算法性能分析,一直到包括多正規化方法、集群算法和並行計算在內的高級主題。因此,這對該領域的研究人員也很有興趣。本書將理論直接與基於Web的計算機代碼相關聯。這使讀者能夠快速開始自己的模擬並輕鬆驗證許多數值示例。目前的代碼是Fortran 77,有免費的編譯器可用。所教授的原則對於使用其他編程語言(如C或C++)的用戶也很重要。

- 取樣、統計和計算機代碼
- 獨立隨機變量的誤差分析
- 馬可夫鏈蒙地卡羅
- 馬可夫鏈數據的誤差分析
- 高級蒙地卡羅
- 並行計算
- 結論、歷史和展望