**Objective**: This series of lectures is an introduction to the broad field of scientific computing. Its main aim is to teach you some tools to decide whether a numerical result produced by a computer code can be trusted or not. Numerical studies are ubiquitous in strategic and political decisions -- such as allowing long-term underground storage of nuclear waste, to name just one example. A computer always gives a numerical result, but is it the correct one?

More precisely, at the end of this series of lectures, you will know

- which questions to ask to assess the reliability of a numerical result, by carefully distinguishing the various possible sources of errors
- some numerical algorithms used on a daily basis (explicitly or implicitly), theirs successes and limitations
- how to implement in a computer code a few basic algorithms and test their limitations

**Programme**:

- general concepts of scientific computing (modelling, error analysis)
- numerical computation of integrals : deterministic and stochastic methods
- numerical integration of ordinary differential equations

**Requirements :**Undergraduate analys and probability theory lectures.

**Evaluation mechanism :**Short reports for the computer hands-on and exercice sessions.

**Last Modification :**Friday 13 July 2012