Particle Methods with Financial Applications


u3500102's picture

u3500102 - Posted on 15 August 2013

Project Description: 

The growing field of Feynman-Kac expectation and related particle models is one of the most active contact points between probability theory and practical applications.

The central theme of the project would be studying how particle methods can be applied in computational finance. Credit risk modeling is of top priority in this regard. Rare event and default probabilities simulation with this state-of-the-art technology will form the major part of the project.

Research Project Details
Project Duration: 
05/2013 to 09/2013
Project Significance: 
The project studies the application of stochastic particle methods in computational finance. Algorithms will be developed to solve high-dimensional problems like multi-name default simulations and parameter estimations, in credit risk analysis. This project is supported by Overseas Research Internship Award, under the Undergraduate Research Fellowship Programme (URFP) 2013 of the University.
Remarks: 
Stochastic particle methods are frequently used to solve highly complex problems. HPC plays a very important role since the particle algorithms can be parallelized easily to approximate a flow of Feynman-Kac measures in high-dimensional problems, such as credit risk analysis of CDOs, the typical dimension of which is 125.