eas@utcsrgv.UUCP (Ann Struthers) (11/01/84)
NUMERICAL ANALYSIS SEMINAR
Thurs. Nov. 8, 1984
3:00 p.m.
"Solution of Premixed and Counterflow Diffusion Flame
Problems by Adaptive Boundary Value Methods"
Professor Mitch Smooke
Department of Computer Science
Yale University
New Haven, Conn. eas@utcsrgv.UUCP (Ann Struthers) (11/02/84)
NUMERICAL ANALYSIS SEMINAR
Thurs. Nov. 6, 1984
3:PM
SANFORD FLEMING BUILDING 1101
"Solution of Premixed and Counterflow Diffusion Flame Problems
by Adaptive Boundary Value Methods"
Professor Mitch Smooke
Department of Computer Science
Yale University
New Haven, Conn.
eas@utcsrgv.UUCP (Ann Struthers) (11/29/84)
NUMERICAL ANALYSIS SEMINAR
Thursday, December 6, 1984
3:00 P.M.
Sandford Fleming Building 1101
Professor Rudi Mathon
Department of Computer Science
University of Toronto
"Boundary Mehtods for Solving Elliptic Problems with
Singularities and Interfaces"
Boundary approximation techniques are described for solving homogeneous
self-adjoint elliptic equations. Piecewise expansions into particular
solutions are used which approximate both the boundary and interface
conditions in a least squares sense. Convergence of such approximations
is proved and error estimates are derived in a natural norm. Numerical
experiments are reported for the singular Motz problem which yield
extremely accurate solutions with only a modest computational effort.voula@utcsri.UUCP (Voula Vanneli) (02/20/85)
University of Toronto
Department of Computer Science
(SF = Sandford Fleming Building, 10 King's College Road)
NUMERICAL ANALYSIS SEMINAR - Thursday, February 28th, 3
p.m., SF 1105
Professor Kenneth Jackson
Dept. of Computer Science, University of Toronto
"Runge-Kutta Methods for the 1990's"
Abstract
Since the development of DVERK by Hull, Enright and Jackson
in the 1970's, there has been a number of advances in formu-
las and heuristics for Runge-Kutta methods. We shall review
some of these advances in general and the development of
interpolants for Runge-Kutta methods in particular.
More specifically, we shall discuss a general procedure for
the construction of interpolants for Runge-Kutta methods.
As illustrations, we shall use this procedure to develop
interpolants for three explicit Runge-Kutta formulas,
including those employed in the well-known subroutines RKF45
and DVERK. A typical result is that only one extra function
evaluation per step is required to obtain a locally sixth-
order interpolant for DVERK.krj@utcsri.UUCP (Ken Jackson) (08/23/85)
Numerical Analysis Seminar Solving Stiff ODEs with Coupled Algebraic Equations on a Microcomputer System Per Grove Thomsen Technical University of Denmark Date: Tuesday, August 27, 1985. Time: 3:10 p.m. Place: Sandford Fleming Building, Room 3207, University of Toronto The typical problem in the design of engineering control systems leads to coupled algebraic and differential equations where the system may change at well-defined set-points. Mathematically, the problem will be that of solving a system of the form y' = f(y,z,t), h(y,z,t) = 0, y(t0) = y0, h(y0,z0,t0) = 0, where y is the vector of dynamic variables and z is the vector of state variables. The set-points are equivalent to points of discontinuity. The implementation is based on SDIRK methods with coupled algebraic equations and special interpolation formulae for smooth continuation of the solution and location of the discontinuities. The programming language is Pascal and the implementation runs on a 64K microcomputer.
wlrush@water.waterloo.edu (Wenchantress Wench Wendall) (04/29/89)
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES
DNUMERICAL ANALYSIS SEMINAR
-Friday, May 5, 1989
Dr. Barry Joe, Dept. of Computing Science, University
of Alberta, will speak on "Construction of 3-D
Triangulations Using Local Transformations."
TIME: 10:30-12:00
ROOM: DC 1302
ABSTRACT
A 2-D (Delaunay) triangulation can be constructed using
a local transformation procedure which swaps the
diagonal edge of 2 adjacent triangles forming a
strictly convex quadrilateral. An analogous local
transformation procedure can be used to construct a 3-D
triangulation, i.e. a connection of n 3-D points into
-
non-overlapping tetrahedrons which fill the convex hull
of the points. This 3-D procedure swaps the interior
faces in 2 or 3 adjacent tetrahedrons forming a convex
hexahedron.
In this talk, we present a new algorithm for
constructing a 3-D Delaunay triangulation using local
transformations. This algorithm has a worst case time
2
complexity of O(n ), which is worst case optimal. For
- -
sets of random points, the expected and empirical time
4/3
-
complexity of this algorithm is O(n ). We also
- -
present a related algorithm in which the local
transformations are not explicitly performed, and
discuss pseudo-locally-optimal non-Delaunay
triangulations and triangulations based on the max-min
solid angle criterion.