Introduction to Numerical Simulation (SMA 5211)
Computational techniques for engineering simulation: sparse and Krylov solvers, Newton methods, ODEs, finite elements and model reduction.
Simulating engineering and physical systems depends on a core set of computational techniques, and this graduate class introduces them. It covers equation formulation, sparse and iterative linear solvers, Newton methods and ODE solvers. It then treats time-periodic problems, molecular dynamics, finite elements, boundary elements, fast multipole methods and model order reduction, with MATLAB files for eight problem sets.
What You Will Achieve
- Formulate network problems with stamping techniques
- Solve large systems using sparse and Krylov methods
- Apply Newton methods to nonlinear problems
- Choose stable methods for ordinary differential equations
Program Overview
Linear Systems
The first lectures formulate and solve equations.
- Equation formulation methods
- Direct methods for sparse linear systems
- QR and Krylov iterative methods
Nonlinear and Time-Dependent Problems
The next lectures add dynamics.
- Multi-D Newton
- ODE solution methods
- Time-periodic solution methods
- Molecular dynamics
Fields and Reduction
The last lectures reach large simulations.
- 3-D elliptic problems
- Finite-element and spectral methods
- Boundary-element approach
- Model order reduction
Course Completion
Job Outlook
These methods are used across engineering, in the design of circuits, structures and materials, and in scientific computing.
Good for: Engineering graduate students with differential equations or linear algebra who are familiar with MATLAB.
Keep in mind: Lecture notes and problem sets are published, but no solutions are listed. It was taught to students at MIT and in Singapore at once, and molecular dynamics uses an external program.
Bottom line: A numerical methods course, with code that gives you a head start on the problems.
Frequently Asked Questions
What software do I need?
Familiarity with MATLAB is expected, and many problem sets supply .m files. One problem set also uses the Moldy molecular dynamics program.
Which engineering fields does it draw on?
Applications come from aerospace, mechanical, electrical, chemical and biological engineering, materials science and operations research.
What supporting files come with the sets?
They include circuit test files, QR code, Poisson and GCR routines, Newton code with test systems, and capacitance calculations on plates and spheres.



There are no reviews yet.