### Session P24: Focus Session: What is Computational Physics? II followed by Computational Methods: Numerical Methods for Strongly Correlated Systems II

 Wednesday, March 23, 2011 8:00AM - 8:12AM P24.00001: US-Japan Workshops on Computational Physics - International Liaison Activities - Yuko Okamoto In this talk, I will report on US-Japan workshops on computational physics. I will summarize what was done in the past workshops that were held in Hawaii and talk about future plans. Wednesday, March 23, 2011 8:12AM - 8:24AM P24.00002: Opportunities for Computational Discovery in Basic Energy Sciences Mark Pederson An overview of the broad-ranging support of computational physics and computational science within the Department of Energy Office of Science will be provided. Computation as the third branch of physics is supported by all six offices (Advanced Scientific Computing, Basic Energy, Biological and Environmental, Fusion Energy, High-Energy Physics, and Nuclear Physics). Support focuses on hardware, software and applications. Most opportunities within the fields of~condensed-matter physics, chemical-physics and materials sciences are supported by the Officeof Basic Energy Science (BES) or through partnerships between BES and the Office for Advanced Scientific Computing. Activities include radiation sciences, catalysis, combustion, materials in extreme environments, energy-storage materials, light-harvesting and photovoltaics, solid-state lighting and superconductivity.~ A summary of two recent reports by the computational materials and chemical communities on the role of computation during the next decade will be provided. ~In addition to materials and chemistry challenges specific to energy sciences, issues identified~include a focus on the role of the domain scientist in integrating, expanding and sustaining applications-oriented capabilities on evolving high-performance computing platforms and on the role of computation in accelerating the development of innovative technologies. ~~ Wednesday, March 23, 2011 8:24AM - 8:36AM P24.00003: Discussion of Opportunities in Computational Physics Bernd Berg , Mark Pederson A discussion of the points raised in the previous talk by Mark Pederson (DOE) will be encouraged, including: (1) Identifying models and strategies for effectively organizing and availing complex computational simulation capabilities to a broader scientific and technical community. (2) Identifying cross-discipline communication of capabilities to ensure sharing of algorithms. (3) Opinions on evolution of overlap between the basic energy scientific mission and the fields that are typically represented by the March-Meeting participants. (4) Interactions between the domains of computational physics, computer science, and applied mathematics. (5) The proper balance between individual and group achievement. (6) What role could DCOMP have in this? Wednesday, March 23, 2011 8:36AM - 8:48AM P24.00004: Enabling Computational Discovery and Design Daryl Hess Advanced cyberinfrastructure(CI), increases in computing power, and increasing use of data volumes are revolutionizing how science is done, changing the nature of the questions we ask, and opening new frontiers.~ From discovering new phenomena and states of matter to the challenge of designing new materials and matter, the focus on problems with many complex interacting degrees of freedom through computational investigation often leads to large amounts of data that require analysis, preservation, curation, and sharing across the community. Data from many sources plays an increasingly important role as a driver of discovery. I will discuss opportunities in computational and data enabled science and in building the CI of the 21$^{st}$ century.~ Sustainable, maintained, and reliable shared software is an important component of a National CI framework that will empower computational scientists to engage the scientific frontiers and the pressing problems around us.~ The success of computational data enabled science requires innovation that leads to paradigms in attacking difficult problems. Education will play an important role in realizing the full potential of computation and data enabled science for discovery and design. Participation of the computational science community is an important ingredient to create a CI that will propel science forward; some self-assembly is required. NSF provides funding opportunities to help. Wednesday, March 23, 2011 8:48AM - 9:00AM P24.00005: Building Foundations for Future Advances in Computational Physics Barry Schneider , Daryl Hess We continue the discussion on laying the cyberinfrastructure foundations to support future advances in computational science. We will focus on how NSF can help and encourage communication with the community in achieving this goal. Wednesday, March 23, 2011 9:00AM - 9:36AM P24.00006: Quo Vadis Computational Physics? Invited Speaker: Richard Scalettar In this Focus Session, and in the preceding DCOMP Invited Symposium, Great Advances of Computational Physics: Past, Present and Future," we have heard about the vision for our field, reviewed (and previewed) cutting-edge numerical work, and considered where the resources might come to support our endeavors. In this talk I will summarize some of the common themes of this discussion, as well as open up the floor for further thoughts on where our APS Division is heading. Wednesday, March 23, 2011 9:36AM - 9:48AM P24.00007: Magnetic impurities in real lattices: A DMRG and ECA study Carlos Busser , George Martins , Khaled Al-Hassanieh , Adrian Feiguin Magnetic interactions between strongly correlated impurities coupled to a sea of conduction electrons is a subject of great interest from both, experimental and theoretical studies. When many magnetic impurities are attached to the same conduction band a rich phase diagram can arise. By one hand the magnetic impurities can be strongly coupled to the spin of the electrons of the conduction band forming a Kondo singlet. By the other, through electron of the conduction band, a spin-spin interaction between the impurities can appears as a consequence the RKKY interaction. A competition between this two singles is expected. For these two effects is important to have a good description of the electrons with energy close to the Fermi level. Systems like the square lattice, with a van-Hove singularity at the middle of the band, or Graphene, with Dirac electrons, or Carbon nanotubes with multiple bands and multiples van Hove singularities need a proper description of the electrons in the lattice Hamiltonian. In this work we present, through a canonical transformation, a numerical method to study problems with several magnetic impurities coupled to arbitrary lattices using DMRG or ECA techniques. Wednesday, March 23, 2011 9:48AM - 10:00AM P24.00008: Quantum Monte Carlo simulations with tensor-network states Jeong Pil Song , R.T. Clay Matrix-product states, generated by the density-matrix renormalization group method, are among the most powerful methods for simulation of quasi-one dimensional quantum systems. Direct application of a matrix-product state representation fails for two dimensional systems, although a number of tensor-network states have been proposed to generalize the concept for two dimensions. We introduce a useful approximate method replacing a 4-index tensor by two matrices in order to contract tensors in two dimensions. We use this formalism as a basis for variational quantum Monte Carlo, optimizing the matrix elements stochastically. We present results on a two dimensional spinless fermion model including nearest- neighbor Coulomb interactions, and determine the critical Coulomb interaction for the charge density wave state by finite size scaling. Wednesday, March 23, 2011 10:00AM - 10:12AM P24.00009: Diagonalization with matrix-product states Chen Liu , Anders Sandvik We consider matrix-product states (MPSs) combined with diagonalization as a method to study correlated quantum many-body systems. The Hamiltonian matrix is constructed in a non-orthogonal basis of MPSs. Diagonalizing this matrix (a generalized eigenvalue problem) gives the ground state as well as excitations. The accuracy is significantly improved compared to individual optimized MPSs. We discuss several ways to generate the MPS basis states in a suitable way and present results for one- and two-dimensional quantum spin systems. Wednesday, March 23, 2011 10:12AM - 10:24AM P24.00010: Multi-scale entanglement renormalization for critical systems Bela Bauer , Liza Huijse , Kareljan Schoutens , Guifre Vidal , Matthias Troyer Determining information about the underlying conformal field theory of a critical system in one dimension, such as the central charge and scaling dimensions, is a notoriously difficult problem for numerical methods. Using a suitable tensor network state, the multi-scale entanglement renormalization ansatz [1], this information is directly accessible [2]. We apply this method to several critical systems in one dimension, including a supersymmetric model for lattice fermions and Yang-Lee chains. \\[4pt] [1] G. Vidal, Phys. Rev. Lett. 99, 220405 (2007)\\[0pt] [2] R.N.C. Pfeifer et al, Phys. Rev. A 79(4), 040301(R) (2009) Wednesday, March 23, 2011 10:24AM - 10:36AM P24.00011: Simulation of fermionic and frustrated lattice models in two dimensions with tensor network algorithms Philippe Corboz , Glen Evenbly , Jacob Jordan , Roman Orus , Guifre Vidal , Bela Bauer , Matthias Troyer , Frederic Mila , Frank Verstraete The simulation of strongly correlated fermionic and frustrated systems in two dimensions is one of the biggest challenges in computational physics. Borrowing ideas and tools from quantum information and condensed matter physics, a new generation of simulation techniques for many-body systems, the so-called tensor network algorithms (e.g. PEPS, MERA), have been proposed in the last few years. These algorithms have been generalized to fermionic systems recently. We present a particularly simple formalism to account for the statistics of fermionic degrees of freedom in a tensor network. Benchmark results confirm the validity of this approach, and show that the computational cost of simulations does not depend a priori on the particle statistics, but on the amount of entanglement in the system. Wednesday, March 23, 2011 10:36AM - 10:48AM P24.00012: Dynamical simulation of integrable and non-integrable models in the Heisenberg picture Dominik Muth , Razmik Unanyan , Michael Fleischhauer The numerical simulation of quantum many-body dynamics is typically limited by the linear growth of entanglement with time. Recently numerical studies have shown, however, that for 1D Bethe-integrable models the simulation of local operators in the Heisenberg picture can be efficient as the corresponding operator-space entanglement grows only logarithmically. Using the spin-1/2 XX chain as generic example of an integrable model that can be mapped to free particles, we here provide a simple explanation for this. We show furthermore that the same reduction of complexity applies to operators that have a high-temperature auto correlation function which decays slower than exponential, i.e., with a power law. This is amongst others the case for models where the Blombergen-De Gennes conjecture of high-temperature diffusive dynamics holds. Thus efficient simulability may already be implied by a single conservation law (like that of total magnetization), as we will illustrate numerically for the spin-1 XXZ model. Wednesday, March 23, 2011 10:48AM - 11:00AM P24.00013: ABSTRACT WITHDRAWN