Diagrammatic Monte Carlo Method for Correlated System

报告人: 邓友金 (中国科学技术大学)

报告时间: 2016年8月19日 15:00

报告地点: 理科楼B315

   报告摘要:Diagrammatic Monte Carlo (DiagMC) method is to stochastically sample the abstract configuration space in terms of perturbative Feynman diagrammatic series. While leading to an exponential growth of computational complexity as diagram order, the fermion sign is also responsible for the convergence of the perturbative diagrammatic series for many quantum systems. DiagMC can be directly applied to the thermodynamic limit—i.e., infinite system size, and moreover, take the advantage of well-known field-theoretical techniques to run the calculation in a self-consistent mode. In this talk, I shall describe the basic formulation of DiagMC, and then discuss two applications. We study the two-dimensional fermionic Hubbard model, provide finite-temperature results and obtain the ground-state phase diagram for superfluids of different symmetries for a significant range of interaction strength and doping fraction. We also find “fingerprint” evidence of the thermal spin-ice state for the quantum Heisenberg antiferromagnet on the pyrochlore lattice—a frustrated system. 

Refs. EPL110 57001 (2015), PRX5 041041 (2016), PRL116 177203 (2016).

   报告人简介:邓友金,男,1974年生于福建。1997年毕业于北京师范大学物理学系,2000年于北京师范大学获得理论物理专业硕士学位,2004年于荷兰Delft University of Technology获得理学博士学位。2009年引进到中国科学技术大学近代物理系。2010年受聘为美国麻省大学Amherst分校兼职助理教授 (Adjunct Assistant Professor)。是Phys. Rev. Lett., Phys. Rev. B, Phys. Rev. E, Phys. Lett. B, Physica A和Physica E的审稿人,也是澳大利亚研究理事会(Australian Research Council, ARC)的评审专家。邓友金教授的研究领域为统计物理、计算物理、和量子模拟。从2000年起,一直从事于高效蒙特卡洛方法的设计及其在统计物理模型的应用,探索相变和临界理论;从2008年开始涉足基于光和冷原子的量子模拟的相关理论;从2010年开始探索强关联量子多体的高效蒙特卡洛方法。自2002年以来共发表研究论文70余篇,其中61篇发表于Nature系列或Physical Review系列。