源凸域正贡献蒙特卡罗方法

在Duoi和Gerstl的贡献蒙特卡罗方法中,不少次级粒子的贡献是负的3,10]。负贡献粒子常使蒙特卡罗估计的统计涨落变大,浪费较多的计算时间,因此,解决次级粒子的负贡献问题是很必要的。在本文中,一个新的称为源凸域正贡献方法的贡献蒙特卡罗方法被提出来了。在任何情况下,源凸域正贡献方法一定比Dubi和Gerstl的贡献方法好,因此,前者完全可以替代后者。对于源凸域R解决了次级粒子的负贡献问题,通过例子的实际计算结果表明,本方法的效率可比Dubi和Gerstl的方法提高1.5至3.4倍。

POSITIVE CONFRIBUTON MONTE CARLO METHOD FOR SOURCE CONVEX REGION

The contribution of more secondaries particles in the contributon Monte Carlo method by Dubi and Gerstl is negative^3,10]. The negatively contributing particle often makes the statistical fluctuation, of the Monte Carlo estimate become large and the computation is more time-consuming. Therefore, if is Very necessary to solve the negatively contributing problem of the secondery particleIn this paper, a new contributon Monte Carlo method called the SCR qositi-ve contributon method is proposed. In any case, the SCR positive contributon method is surely betfer than the contribufon method by Dubi and Gerstl hence the former can replace the latter thorughly, too. The negativily contributing prob-lem of the secondary particle for tho source convex region is solved, The results of the practical calculation by examples indicates that the efficiency of the present method can be raised by 1.5 of 3.4 times in comparison with the method by Dubi and Gerstl.