2nd European Workshop on Automatic Differentiation

Thursday November 17- Friday November 18, 2005

Whitworth Conference Centre

Cranfield University (Shrivenham Campus)

Defence Academy of the UK

Shrivenham, Swindon

Des Higham (University of Strathclyde)

Opportunities for AD in Stochastic DE Solving

Abstract: I will give a brief and accessible introduction to numerical methods for simulating stochastic differential equations. This is an area where very little general purpose software is available and AD is an attractive way forward. On one hand I will argue that AD has clear potential for the automatic generation of high-order methods, via repeated application of Ito's lemma. On the other hand for low-order and/or implicit methods, AD can still overcome inherent difficulties associated with finite differences and frozen Jacobians due to the lack of smoothness in solution paths.

Slides: HighamNov05.pdf