非定常粒子输运蒙特卡罗散射源分层抽样方法

计算物理 ›› 2005, Vol. 22 ›› Issue (6): 57-63.

非定常粒子输运蒙特卡罗散射源分层抽样方法

邓力¹, 张文勇², 黄正丰¹, 王瑞宏¹, 许海燕¹, 李树¹

1. 北京应用物理与计算数学研究所计算物理实验室, 北京 100088;
2. 国防科技大学计算机学院, 湖南长沙 410073

收稿日期:2004-09-10 修回日期:2005-02-17 出版日期:2005-11-25 发布日期:2005-11-25
作者简介:邓力(1960-), male,Mianzhu,Sichuan,PhD Professor,Monte Carlo method and its plication.
基金资助:
supported by the unit foundation of National Nature Science Foundation Committee of China and Chinese Academy of Engineering Physics and National Key Laboratory Computational Physics

A Stratified Sample Method of Scattering Source for Time-dependent Monte Carlo Transport

DENG Li¹, ZHANG Wen-yong², HUANG Zheng-feng¹, WANG Rui-hong¹, XU Hai-yan¹, LI Shu¹

1. Lab. Com. Phys., Institute of Applied Physics and Computational Mathematics, Beijing 100088, China;
2. Computer Institute, National University of Defense Technology, Changsha 410073, China

Received:2004-09-10 Revised:2005-02-17 Online:2005-11-25 Published:2005-11-25
Supported by:
supported by the unit foundation of National Nature Science Foundation Committee of China and Chinese Academy of Engineering Physics and National Key Laboratory Computational Physics

摘要/Abstract

摘要： 定常粒子输运蒙特卡罗并行计算是成功的,因为粒子游动是独立的,可以把模拟的粒子数等分到每个处理器去.然而,对非定常问题,由于每个时间步涉及散射源和几何网格的通讯,它严重的制约了并行规模,导致并行不可扩展.研究了两种算法,采用自适应分配处理器,提高了加速比和处理器的利用率;采用蒙特卡罗分层抽样大大降低了处理器之间散射源的通讯量,并行可扩展性显著改善,取得了理想的加速比.

关键词: 非定常, 蒙特卡罗输运, 自适应处理器分配, 散射源分层抽样

Abstract: A parallel algorithm for time-independent Monte Carlo transport is successful since particles are independent and they are distributed to multiple processors.However,for time-dependent Monte Carlo transport problems, the parallel efficiency reduces and the parallel scale is limited due to the communication of scattering source attribute and meshes in each time-step.We propose two algorithms in them adaptive processor assignment and optimized processor choice are obtained.With a Monte Carlo stratified sampling technique for scattering source treatment the communication cost is reduced greatly.The parallel expandability is improved.A large speedup over the basic algorithm is obtained.

Key words: time-dependent, Monte Carlo transport, adaptive processor assignment, stratified sampling of scattering source

中图分类号:

O571.51

邓力, 张文勇, 黄正丰, 王瑞宏, 许海燕, 李树. 非定常粒子输运蒙特卡罗散射源分层抽样方法[J]. 计算物理, 2005, 22(6): 57-63.

DENG Li, ZHANG Wen-yong, HUANG Zheng-feng, WANG Rui-hong, XU Hai-yan, LI Shu. A Stratified Sample Method of Scattering Source for Time-dependent Monte Carlo Transport[J]. CHINESE JOURNAL OF COMPUTATIONAL PHYSICS, 2005, 22(6): 57-63.

[1]	上官丹骅, 许海燕. 非定常输运模拟中基于粒子标识分类的源偏倚算法[J]. 计算物理, 2016, 33(6): 639-644.
[2]	刘祥, 孙秦, 武亮. 一种基于连续核函数展开的跨声速气动力修正法[J]. 计算物理, 2015, 32(4): 416-422.
[3]	孙晨扬, 李启良, 杨志刚. 最小二乘有限元法求解非定常应力的Navier-Stokes方程[J]. 计算物理, 2015, 32(1): 13-19.
[4]	周春华. 一种基于预估-修正策略的非定常流网格自适应加密/稀疏方法[J]. 计算物理, 2013, 30(5): 633-641.
[5]	赵家权, 刘培启, 刘凤霞, 赵文静, 王亮, 胡大鹏. 激波管中非平衡可凝结流的流动分析[J]. 计算物理, 2012, 29(5): 707-712.
[6]	马小亮, 杨国伟. 凹腔非定常特性的数值模拟[J]. 计算物理, 2010, 27(3): 375-380.
[7]	吕俊明, 胡宗民, 姜宗林. 亚声速段横向射流对Coil性能影响的数值研究[J]. 计算物理, 2008, 25(4): 414-420.
[8]	赵慧勇, 乐嘉陵. 双时间步方法的应用分析[J]. 计算物理, 2008, 25(3): 253-258.
[9]	欧平, 马汉东, 汪翼云. 任意马赫数非定常流动数值模拟的统一算法[J]. 计算物理, 2007, 24(2): 166-170.
[10]	杨云军, 周伟江, 崔尔杰. 耦合时间精度对模拟飞行器自由运动特性的影响[J]. 计算物理, 2007, 24(1): 42-48.
[11]	薛具奎. 求解非定常不可压N-S方程的预处理方法[J]. 计算物理, 2002, 19(5): 403-407.
[12]	马汉东, 袁宏生. 非定常喷流有限差分方法数值试验[J]. 计算物理, 2001, 18(4): 377-380.
[13]	秦国良, 徐忠. 谱元方法求解正方形封闭空腔内的自然对流换热[J]. 计算物理, 2001, 18(2): 119-124.
[14]	钟方平. 柱形容器内部爆炸流场的数值模拟[J]. 计算物理, 2000, 17(6): 695-701.
[15]	郭加宏, 刘高联. 机翼跨声速非定常绕流I_A型反命题变域变分有限元解[J]. 计算物理, 2000, 17(5): 518-524.