How to start knowledge build-up for becoming a numerical modeler?

[paul143]

Hi everyone,

I want to know what disciplines must I study to become an effective numerical modeler?
As a numerical modeler, I expect to be able to generate numerical solutions to mathematical problems (via programming) at the same time I want to be able to explain all aspects of my model/simulation results and other aspects of the process involved.

I am fluent in fortran90, python and java, but have never been able to apply these acquired skills in solving numerical plroblems.

Any suggestions?

 

[Lavabug]

Take a numerical methods course or two intended for physicists/engineers.

 

[MathWarrior]

Scientific Computing would be a good course. Also check out coursera.org it has a free class on scientific computing I think.

 

[Lavabug]

These course notes are pretty good, a large part of it is what I covered in my computational physics/numerical methods course:

http://ocw.mit.edu/courses/mechanical-engineering/2-29-numerical-fluid-mechanics-spring-2007/lecture-notes/

 

[paul143]

Thanks for the suggestion guys! Any books that you suggest I should start reading? Would it be best for me to also learn C? Matlab?

 

[Lavabug]

Learn how to code, forget about the language. If you already know f90 and python you won't have trouble translating MATLAB course exercises to them (but there exist several books on the subject that use those languages).

Just stick to the exercises in whatever book you use, there are a lot of "semester course size" books that have lots of example code (often matlab) and then there are bigger reference texts that are more complete (Schwartz "Numerical Analysis", rigorous but only has pseudocode.) Just head over to your library and find one you like.

 

[chiro]

Hey paul143.

One good set of books for this is the Numerical Recipe Book series:

http://apps.nrbook.com/c/index.html