Think about a situation, where a financial institution has many option positions, each written on a different underlying asset, and there is an unexpected arrival of market-wide news that shakes the markets.
Reliable option pricing and hedging will require an instant re-calibration of the volatility models on the different underlyings. If the organization uses advanced models, the calibration of models using data on multiple underlying securities can take too long, and will put the business under risk.
Prof. Juho Kanniainen has published a paper that demonstrates how the calibration work can be accelerated and done in near real time using Techila Distributed Computing Engine. In his work, he shows how to cut the computing time from more than 7 hours to just a couple of minutes. Prof. Kanniainen has made his MATLAB code open and available for the financial community.
Download Prof. Kanniainen’s MATLAB code here.