Speakers
Prof.
Tetsuya Takaishi
(Hiroshima University of Economics)Prof.
Yubin Liu
(Nankai University)
Description
The hybrid Monte Carlo (HMC) algorithm has been widely used for dynamical lattice QCD simulations. One of the advantages of using the HMC algorithm is that it is a global algorithm that can update all link variables simultaneously. In this way we can greatly reduce computational cost concerning the fermionic part. We utilize this advantage for parameter estimations of the realized stochastic volatility model which has been used for modelling time series data. The realized stochastic volatility model includes a number of volatility variables to be updated. We update those variables by the HMC algorithm. It is found that the HMC algorithm de-correlates effectively Monte Carlo samples of volatility variables. We also show that the algorithm can be accelerated by the GPU computing.
Primary author
Prof.
Tetsuya Takaishi
(Hiroshima University of Economics)
Co-authors
Dr
Tingting Chen
(Hiroshima University)
Prof.
Yubin Liu
(Nankai University)
