A Review of Common Random Numbers in Simulation

Beschreibung

This paper presents an analysis of the Common Random Numbers methodology, including its theoretical foundation, advantages and disadvantages, mathematical formulations, and implementation in R.
Common Random Numbers (CRN) are a core technique in the toolbox of variance reduction methods used in simulation. The central idea behind CRN is to use the same sequence of random numbers across different system simulations to induce correlation that, under certain conditions, leads to variance reduction. This paper presents an analysis of the CRN methodology, including its theoretical foundation, practical implementation, potential advantages and disadvantages, mathematical formulations, and implementation in R. Two straightforward simulations illustrate cases where CRN dramatically reduces variance and where it counterproductively increases it.

Autor*in

Johann Markus Schauerhuber
During his studies and academic resp. professional activities Prof. Dr. Dr. Johann Markus Schauerhuber has been intensively involved in statistical programming, stochastic, risk theory, simulation and mathematical modeling. Most of his professional experience has been gained in the university domain as an academic director, postdoc lecturer / researcher and in government authorities. Email: jm_schauerhuber@gmx.at

Themen in »A Review of Common Random Numbers in Simulation«

Common Random Numbers Simulation Variance Reduction

Stimmen zu »A Review of Common Random Numbers in Simulation«

Details

Link teilen