项目名称: 蠕虫类蒙特卡洛算法的发展和应用
项目编号: No.11275185
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 数理科学和化学
项目作者: 邓友金
作者单位: 中国科学技术大学
项目金额: 80万元
中文摘要: 蒙特卡洛方法是统计和凝聚态物理研究中一个不可缺少的工具。常见的高效算法有集团和蠕虫算法。通过借鉴已有的各种技术和数据结构,本项目致力于设计新型高效的蠕虫类算法,并应用在一些典型的经典和量子多体系统。首先,作为以前项目的延续和拓展,研究在现代相变理论和场论占有重要地位的Potts反铁磁和圈模型;结合解析映射、赋色、链接搜寻、以及小概率处理等技术,设计高效蠕虫类算法;探索系统的相变性质。其次,设计路径积分表象下连续虚时蠕虫算法(特别是有四体相互作用),研究玻色混合体、几何阻挫晶格上的硬核玻色系统、以及J-Q自旋模型,探索超固态等新奇宏观量子现象和检验解禁闭相变理论的正确性。最后, 也最重要的,应用完全粗化、数值戴森方程、自设计的赝势、蠕虫更新、及自洽思想等,发展费曼图图形蒙特卡洛方法。基于促进收敛的符号祝福,我们预测能首次探测到高维费米Hubbard模型的部分相图及其物理性质。
中文关键词: 蒙特卡洛蠕虫算法;玻色哈勃模型;冷原子;逾渗;相图
英文摘要: Monte Carlo method is an indispensable research tool in statistical and condensed-matter physics. Among the highly efficient ones are cluster and worm algorithms. In this project, by borrowing important physical ideas from the exisiting algorithms and making use of special data structures, we aim to design and develop novel efficient Monte Carlo methods based on worm-type update. First,as a continuation and extension of our earlier research, we shall study the Potts antiferromagnet and the loop model, two important models in the modern theory of phase transition and the quantum field theory. The newly designed worm-type algorithm will incorporate a number of techniques including analytical transformation, coloring, connectivity checking, and cumulative probability for rare events etc. Next, we shall develop the continuous-time worm algorithm for the path-integral representation of some bosonic and spin systems, including multi-component bosonic mixture,hard-core bosons on geometrically frustrated lattices, and the J-Q Heisenberg model of four-body interaction. This is to observe emergent macroscopic quantum phenomena like supersolidity, explore the underlying physical mechanism, and confirm/falsify the deconfined quantum critical theory. Last but not least, we shall develop bold diagrammatic Monte Carlo method f
英文关键词: Monte Carlo worm algorithm;Bose-Hubbard model;cold atom;percolation;phase diagram