Steven J. Owen - Sandia National Laboratories

Title: An Introduction to Mesh Generation Algorithms

Abstract: This talk is a brief introduction to some of the fundamental algorithms used in commercial mesh generation tools. It will cover triangle, tetrahedral, quadrilateral, hexahedral as well as hex-dominant approaches. Delaunay, Advancing Front and Octree approaches will be discussed with respect to triangle and tetrahedral methods. Quad and hex methods will include mapping, submapping, sweeping, paving, q-morph, plastering, h-morph as well as an introduction to selected research oriented methods. An introduction to 3D and parametric surface meshing methods will also be provided. A classification and comparison of existing mesh generation methods will be discussed, showing strengths and weaknesses for various applications. This course is intended to be an introductory course for those new to the field or who would like a non-technical refresher course on basic mesh generation algorithms.

Biography: Dr. Steve Owen is employed by Sandia National Laboratory in Albuquerque, New Mexico and is the current project lead and principal investigator for the CUBIT Geometry and Mesh Generation Toolkit. Past work has focused on facet-based geometry representations for mesh generation, unstructured quadrilateral and hexahedral algorithms, parametric surface meshing, boundary layer meshing for CFD, Delaunay methods, smoothing and topology cleanup, mesh sizing control, among others. He has extensive publication and editorial experience in the mesh generation community and maintains the Meshing Research Corner web site. Prior to Sandia, Steve worked in industry at Ansys Inc., a commercial finite element analysis company based in Pittsburgh Pennsylvania, where he developed and maintained mesh generation tools for commercial use. Steve received his Ph.D. from Carnegie Mellon University in 1999 while working for Ansys Inc. and received his Bachelors and Masters degrees from Brigham Young University in 1992. He currently serves on the graduate committee for several students at CMU and BYU.

Mark S. Shephard - Rensselaer Polytechnic Institute

Title: Reliable Automation of Large-Scale Simulations

Abstract: The reliable automation of simulations requires the effective combination of generalized simulation specification, with the procedures that can automatically construct and adaptively control the simulation models. The reliable simulation of large-scale problems further requires that all of these procedures operate on passively parallel computers. This short course lecture will consider the current state of the technologies needed to meet these goals when the simulations are to be performed over general 3-D domains and employ multiple physical models that can act over multiple scales. Two main areas will be stressed in the presentation. The first will be a computational and software infrastructure to support adaptive multimodel simulations. The second will be the execution and adaptive control of unstructured mesh simulations on massively parallel computers.  Example applications will be given in both areas.

Biography: Mark S. Shephard is the Samuel A. and Elisabeth C. Johnson, Jr. Professor of Engineering, and the director of the Scientific Computation Research Center at Rensselaer Polytechnic Institute. He holds joint appointments in the departments of Mechanical, Aerospace and Nuclear Engineering; Civil and Environmental Engineering; and Computer Science. Dr. Shephard has published over 250 papers. He is a fellow in and the past President of the US Association for Computational Mechanics, a fellow and member of the General Council of the International Association for Computational Mechanics, a fellow of ASME and an Associate Fellow of AIAA. He is the editor of Engineering with Computers and on the editorial board of six computational mechanics journals. He is a co-founder of Simmetrix Inc., a company dedicated to the technologies that enable simulation-based engineering.

Kenji Shimada - Carnegie Mellon University

Title: Current Trends and Issues in Automatic Mesh Generation

Abstract: This tutorial presents current trends and issues in automatic mesh generation. Although automated mesh generation methods in two and three dimensions have been studied intensively, many analysis engineers still craft meshes manually for a certain class of analysis problems. In order to realize fully automated high-quality mesh generation, two technical issues need to be addressed: (1) automated mesh generators should be able to control the anisotropy and directionality of a mesh, and (2) geometric operations required prior to mesh generation should be made more robust and automated. This tutorial outlines recent development of the two technical issues in order to encourage further research and development of advanced mesh generation technology.

Biography: Kenji Shimada is Theodore Ahrens Professor in Engineering at Carnegie Mellon University in the Department of Mechanical Engineering, Biomedical Engineering (by courtesy), Civil and Environmental Engineering (by courtesy) and the Robotics Institute (by courtesy). Dr. Shimada received his B.S. and M.S. from the University of Tokyo, and his Ph.D. from the Massachusetts Institute of Technology. His research interests are in the areas of geometric modeling, mesh processing, computer graphics, medical imaging, and robotics. Prior to joining Carnegie Mellon in 1996, he was Manager of Graphics Applications at IBM Research. Dr. Shimada received the JSAIM Best Author Award in 2006, the ASME Design Automation Best Paper Award in 2004, the IPSJ Best Paper Award in 2002, NSF CAREER Award in 2000, Honda Initiation Grant in 1998, the IPSJ Yamashita Award in 1994, and the Nicograph Best Paper Award in 1994. He is a member of ACM, ASME, IEEE, JSIAM, and SAE.

Nikos Chrisochoides - The College of William and Mary

Title: Parallel Mesh Generation

Abstract: Parallel mesh generation is a  relatively new research area transcending the boundaries of  two scientific computing  disciplines: computational geometry and  parallel computing.  In this tutorial we will present both the theoretical foundation and the practical aspects related to the  implementation of parallel  mesh generation methods on current and  emerging architectures. Parallel  mesh generation methods decompose the original mesh  generation problem  into    smaller subproblems  which  are   solved in  parallel.  We  will  organize the parallel mesh generation methods in terms of two basic attributes: (1) the sequential techniques used  for meshing the individual subproblems and  (2) the  degree  of coupling   between  the subproblems. We  will briefly describe the well known sequential methods and identify common abstractions used in their parallelization.

The goals of the  tutorial are: (1) to  overview the existing parallel mesh generation  methods and derive a  general framework, (2) to apply this framework  to the  analysis  and the  parallelization of Delaunay
based   methods, and  (3) to    promote an  off-line discussion  about specific sequential methods from the audience. The target audience is: (1) graduate students from engineering, computer and applied sciences,
(2)    engineers from industry   who are   interested  to re-train  in parallel, real-time computing and  mesh generation. The  content level is 25% beginners,  50% intermediate, and  25%  advanced. There are  no pre-requisites, however the  attendees  are encouraged to  bring their own methods for off-line discussion at the end of the tutorial.

The topics we will cover are: (1) introduction  to widelyused sequential mesh  generation methods (general definitions and notation) and their  classification into  refinement- and  tiling-based methods;
(2) model  decomposition using both  domain- and data-centric methods; (3) parallel  mesh   generation methodology  and  classification  into decoupled, partially- and  tightly-coupled  methods; (4) a  case study
using Delaunay-based methods; (5)  parallel implementation and runtime software support systems  for mesh  generation  methods; (6)  parallel out-of-core techniques for very large size problems.

Biography: Nikos Chrisochoides is a full time Professor in the Computer Science Dept., Director for the Center for Real-Time Computing and John Simon Guggenheim Fellow in Medicine and Health.

His research interests are in medical image computing and parallel and distributed scientific computing. Specifically, parallel mesh generation both theoretical and implemtation aspects. His research is application-driven. Currently his is working on real-time mesh generation for biomedical applications like non-rigid registration for Image Guided Neurosurgery.

Chrisochoides received his BS in Mathematics from Aristotle University in Greece, his MS (in Mathematics) and PhD (in Computer Science) degrees from Purdue University. Then he moved to Northeast Parallel Architectures Center (NPAC) at Syracuse University as the Alex Nason Postdoctoral Fellow in Computational Sciences. After NPAC he worked in the Advanced Computing Research Institute, at Cornell University . He joined (as an Assistant Professor in January 1997) the Dept. of Computer Science and Engineering at the University of Notre Dame. In the Fall of 2000, he moved to the College of William and Mary as an Associate Professor and in 2004 he was awarded the Alumni Memorial Distringuished Professorship. Chrisochoides has more than 120 technical publications in parallel scientific computing.

He has held visiting positions at Harvard Medical School (Spring 2005), MIT (Spring 2005), Brown University (Fall 2004), and NASA/Langley (Summer 1994).

Andrey Chernikov - The College of William and Mary

Title: Parallel Mesh Generation

Abstract: Parallel mesh generation is a  relatively new research area transcending the boundaries of  two scientific computing  disciplines: computational geometry and  parallel computing.  In this tutorial we will present both the theoretical foundation and the practical aspects related to the  implementation of parallel  mesh generation methods on current and  emerging architectures. Parallel  mesh generation methods decompose the original mesh  generation problem  into    smaller subproblems  which  are   solved in  parallel.  We  will  organize the parallel mesh generation methods in terms of two basic attributes: (1) the sequential techniques used  for meshing the individual subproblems and  (2) the  degree  of coupling   between  the subproblems. We  will briefly describe the well known sequential methods and identify common abstractions used in their parallelization.

The goals of the  tutorial are: (1) to  overview the existing parallel mesh generation  methods and derive a  general framework, (2) to apply this framework  to the  analysis  and the  parallelization of Delaunay
based   methods, and  (3) to    promote an  off-line discussion  about specific sequential methods from the audience. The target audience is: (1) graduate students from engineering, computer and applied sciences,
(2)    engineers from industry   who are   interested  to re-train  in parallel, real-time computing and  mesh generation. The  content level is 25% beginners,  50% intermediate, and  25%  advanced. There are  no pre-requisites, however the  attendees  are encouraged to  bring their own methods for off-line discussion at the end of the tutorial.

The topics we will cover are: (1) introduction  to widelyused sequential mesh  generation methods (general definitions and notation) and their  classification into  refinement- and  tiling-based methods;
(2) model  decomposition using both  domain- and data-centric methods; (3) parallel  mesh   generation methodology  and  classification  into decoupled, partially- and  tightly-coupled  methods; (4) a  case study
using Delaunay-based methods; (5)  parallel implementation and runtime software support systems  for mesh  generation  methods; (6)  parallel out-of-core techniques for very large size problems.

Biography: Andrey Chernikov is a Postdoctoral Research Associate in Computer Science and Center for Real-Time Computing at the College of William and Mary. He graduated with a Ph.D. in Computer Science from the College of William and Mary in 2008 and received a Distinguished Dissertation Award. He received his Masters (2001) and Bachelors (1999) degrees with Distinction in Applied Mathematics and Computer Science from Kabardino-Balkar State University in Russia. In 2006 he held a visiting research position at Brown University, and from 1997 to 2001 he worked at the Instituted of Informatics and Problems of Regional Management of Russian Academy of Sciences in Nalchik, Russia. His research interests include Computational Geometry and specifically mesh generation, Parallel and Distributed Computing, and the development of scientific computing software.