본문

서브메뉴

Towards More Efficient Stochastic Methods in Quantum Chemistry
Towards More Efficient Stochastic Methods in Quantum Chemistry
Towards More Efficient Stochastic Methods in Quantum Chemistry

Detailed Information

Material Type  
 단행본
 
0017162283
Date and Time of Latest Transaction  
20250211151954
ISBN  
9798384048060
DDC  
530
Author  
Anderson, Tyler Axel.
Title/Author  
Towards More Efficient Stochastic Methods in Quantum Chemistry
Publish Info  
[Sl] : Cornell University, 2024
Publish Info  
Ann Arbor : ProQuest Dissertations & Theses, 2024
Material Info  
153 p
General Note  
Source: Dissertations Abstracts International, Volume: 86-03, Section: B.
General Note  
Advisor: Umrigar, Cyrus.
학위논문주기  
Thesis (Ph.D.)--Cornell University, 2024.
Abstracts/Etc  
요약The challenge of determining the ground state energy in quantum chemistry has led to the development of many computational methods, each with their own advantages and disadvantages. In this dissertation, we focus on two such methods - diffusion Monte Carlo (DMC) and the stochastic heatbath configuration interaction (SHCI) method. Both methods are systematically improvable and therefore can, at least in principle, be used to determine the exact energy of any system given enough computational resources. In practice, various approximations must be made to keep the computational cost under control, and determining the exact energy is only possible after controlling for various errors and extrapolating to various limits. We develop several improvements to each method which aid in this process of determining the true energy of chemical systems.In Chapter 2 we focus on DMC. Nonlocal pseudopotentials are often used to remove core electrons to improve computational efficiency. For these systems, we introduce an additional Metropolis-Hastings accept-reject step in the T-moves algorithm which results in lower time-step errors in the total energy and a greatly improved distribution sampled by DMC. The time-step errors in observables which do not commute with the Hamiltonian are especially improved. We also introduce an exact expression for the nonlocal part of the Green's function, for which a linear approximation in the time-step is often used. This leads to only minor improvements, but is straightforward to implement. We then introduce a number of criteria which the ideal reweighting factor should have, and introduce an example of a reweighting factor which meets all criteria. Using this reweighting factor, we show that the time-step error in the total energies of both pseudopotential and all-electron calculations are consistently improved for a variety of systems.In Chapter 3 we present improvements to SHCI. First, we introduce an array called the singles queue which allows us to improve the efficiency of selecting important singly excited Slater determinants. Each element of this array contains an upper bound on the absolute value of the Hamiltonian matrix element between any single excitation involving a given pair of orbitals, which allows us to very efficiently eliminate single excitations which cannot satisfy the SHCI criterion. Second, we introduce a basis-set correction based on range-separated density functional theory (DFT) which aids extrapolation to the complete basis set (CBS) limit. Because SHCI, in common with all other quantum chemistry methods except DMC which works directly in the complete basis limit, typically converges very slowly with respect to the size of the basis, this improvement greatly improves our ability to reach chemical accuracy with SHCI.
Subject Added Entry-Topical Term  
Physics
Subject Added Entry-Topical Term  
Quantum physics
Subject Added Entry-Topical Term  
Computational physics
Index Term-Uncontrolled  
Diffusion Monte Carlo
Index Term-Uncontrolled  
Density functional theory
Index Term-Uncontrolled  
Quantum chemistry
Index Term-Uncontrolled  
Computational methods
Added Entry-Corporate Name  
Cornell University Physics
Host Item Entry  
Dissertations Abstracts International. 86-03B.
Electronic Location and Access  
로그인 후 원문을 볼 수 있습니다.

MARC

 008250123s2024        us                              c    eng  d
■001000017162283
■00520250211151954
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798384048060
■035    ▼a(MiAaPQ)AAI31328956
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a530
■1001  ▼aAnderson,  Tyler  Axel.▼0(orcid)0000-0002-0276-8860
■24510▼aTowards  More  Efficient  Stochastic  Methods  in  Quantum  Chemistry
■260    ▼a[Sl]▼bCornell  University▼c2024
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2024
■300    ▼a153  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  86-03,  Section:  B.
■500    ▼aAdvisor:  Umrigar,  Cyrus.
■5021  ▼aThesis  (Ph.D.)--Cornell  University,  2024.
■520    ▼aThe  challenge  of  determining  the  ground  state  energy  in  quantum  chemistry  has  led  to  the  development  of  many  computational  methods,  each  with  their  own  advantages  and  disadvantages.  In  this  dissertation,  we  focus  on  two  such  methods  -  diffusion  Monte  Carlo  (DMC)  and  the  stochastic  heatbath  configuration  interaction  (SHCI)  method.  Both  methods  are  systematically  improvable  and  therefore  can,  at  least  in  principle,  be  used  to  determine  the  exact  energy  of  any  system  given  enough  computational  resources.  In  practice,  various  approximations  must  be  made  to  keep  the  computational  cost  under  control,  and  determining  the  exact  energy  is  only  possible  after  controlling  for  various  errors  and  extrapolating  to  various  limits.  We  develop  several  improvements  to  each  method  which  aid  in  this  process  of  determining  the  true  energy  of  chemical  systems.In  Chapter  2  we  focus  on  DMC.  Nonlocal  pseudopotentials  are  often  used  to  remove  core  electrons  to  improve  computational  efficiency.  For  these  systems,  we  introduce  an  additional  Metropolis-Hastings  accept-reject  step  in  the  T-moves  algorithm  which  results  in  lower  time-step  errors  in  the  total  energy  and  a  greatly  improved  distribution  sampled  by  DMC.  The  time-step  errors  in  observables  which  do  not  commute  with  the  Hamiltonian  are  especially  improved.  We  also  introduce  an  exact  expression  for  the  nonlocal  part  of  the  Green's  function,  for  which  a  linear  approximation  in  the  time-step  is  often  used.  This  leads  to  only  minor  improvements,  but  is  straightforward  to  implement.  We  then  introduce  a  number  of  criteria  which  the  ideal  reweighting  factor  should  have,  and  introduce  an  example  of  a  reweighting  factor  which  meets  all  criteria.  Using  this  reweighting  factor,  we  show  that  the  time-step  error  in  the  total  energies  of  both  pseudopotential  and  all-electron  calculations  are  consistently  improved  for  a  variety  of  systems.In  Chapter  3  we  present  improvements  to  SHCI.  First,  we  introduce  an  array  called  the  singles  queue  which  allows  us  to  improve  the  efficiency  of  selecting  important  singly  excited  Slater  determinants.  Each  element  of  this  array  contains  an  upper  bound  on  the  absolute  value  of  the  Hamiltonian  matrix  element  between  any  single  excitation  involving  a  given  pair  of  orbitals,  which  allows  us  to  very  efficiently  eliminate  single  excitations  which  cannot  satisfy  the  SHCI  criterion.  Second,  we  introduce  a  basis-set  correction  based  on  range-separated  density  functional  theory  (DFT)  which  aids  extrapolation  to  the  complete  basis  set  (CBS)  limit.  Because  SHCI,  in  common  with  all  other  quantum  chemistry  methods  except  DMC  which  works  directly  in  the  complete  basis  limit,  typically  converges  very  slowly  with  respect  to  the  size  of  the  basis,  this  improvement  greatly  improves  our  ability  to  reach  chemical  accuracy  with  SHCI.
■590    ▼aSchool  code:  0058.
■650  4▼aPhysics
■650  4▼aQuantum  physics
■650  4▼aComputational  physics
■653    ▼aDiffusion  Monte  Carlo
■653    ▼aDensity  functional  theory
■653    ▼aQuantum  chemistry
■653    ▼aComputational  methods
■690    ▼a0605
■690    ▼a0599
■690    ▼a0216
■71020▼aCornell  University▼bPhysics.
■7730  ▼tDissertations  Abstracts  International▼g86-03B.
■790    ▼a0058
■791    ▼aPh.D.
■792    ▼a2024
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17162283▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

Preview

Export

ChatGPT Discussion

AI Recommended Related Books


    New Books MORE
    Statistics for the past 3 years. Go to brief

    Detail Info.

    • Reservation
    • Not Exist
    • My Folder
    • First Request
    • Non-Book Loan Application
    • Nighttime Book Loan Application
    Material
    Reg No. Call No. Location Status Lend Info
    TF13655 전자도서 대출가능 My Folder 부재도서신고 비도서대출신청 야간 도서대출신청

    * Reservations are available in the borrowing book. To make reservations, Please click the reservation button

    Books borrowed together with this book

    Related Popular Books

    Available after logging in.