統計変動を抑えるための整乱数

概要

論文の詳細を見る
Statistical errors associated with counting are a common issue in Monte Carlo simulations. Increasing the trial frequency suppresses the statistical errors associated with counting and improves the reliability of sampling results. This leads to calculation times that increase to unacceptable levels for large-scale simulations. In a practical simulation, it is necessary to obtain "results with reliability" with a minimum number of trials. Various random number distributions are used for Monte Carlo simulation. A method called "refined random number" generation has been developed whereby the "statistical counting error" is adjusted for "Various random numbers." In this paper, the method for generating refined random numbers and the consequent enhanced reliability simulations using refined random numbers are described.
北海道工業大学の論文
2007-03-26