Supercomputer simulation of ITO stochastic differential equations. Phase 1
It was the intent of the project to improve the methods for simulating stochastic differential equations, both in terms of optimizing computer codes for execution time and improving the accuracy of determining functionals of the simulated process. By appropriate modifications of the existing theory about stability in numerical methods for ordinary differential equations, researchers found encouraging results for stochastic differential equations. They concluded that sample sizes and stability are far more important than the step size for obtaining satisfactory results. The methods are well adapted for vector or parallel computers. Examining vector/parallel computing methods for such processes inspires new methods for generating uniform and gaussian random numbers and produces a new and welcome attitude change about such simulations.
- Research Organization:
- Principia Supervectus, Inc., Seattle, WA (USA)
- OSTI ID:
- 6597247
- Report Number(s):
- PB-90-257387/XAB; CNN: NSF-ISI8660377
- Country of Publication:
- United States
- Language:
- English
Similar Records
The fully implicit stochastic-{alpha} method for stiff stochastic differential equations
Stabilized multilevel Monte Carlo method for stiff stochastic differential equations
Related Subjects
990200* -- Mathematics & Computers
BROWNIAN MOVEMENT
COMPUTERIZED SIMULATION
COMPUTERS
DIFFERENTIAL EQUATIONS
DIGITAL COMPUTERS
EQUATIONS
FORTRAN
GAUSSIAN PROCESSES
NONLINEAR PROBLEMS
NUMERICAL SOLUTION
OPTIMIZATION
PARALLEL PROCESSING
PROGRAMMING
PROGRAMMING LANGUAGES
RANDOM NUMBER GENERATORS
SIMULATION
STOCHASTIC PROCESSES
SUPERCOMPUTERS
VECTOR PROCESSING