Multivariate Random Numbers in a Monte Carlo Simulation

John Flood

2015 4 24

0 个回答

13 次查看（30 天）

显示更早的评论

0 个投票

Good day

I am doing a Monte Carlo simulation for my dissertation study. I would like to generate multivariate random numbers for my study but I am very new to Matlab and I am not entirely sure on how to go about it?

I will be using three asset classes in my study and therefore will need to generate random numbers for each asset class (Equity, Bonds and Bills). From my understanding of generating multivariate random numbers in Matlab the formula should look like this:

R = mvnrnd([mean.equity, mean.bonds, mean.bills], [cov.equity.equity, cov.equity.bonds, cov.equity.bills; cov.bonds.equity, cov.bonds.bonds, cov.bonds.bills; cov.bills.equity, cov.bills.bonds, cov.bills.bills], iterations)

The specific figures used in my study are as follows:

Following the the above function it should look something like this:

R = mvnrnd([.1275,.0688,.06],[ 0.05301539, 0.00000591, 0.00001438; 0.00000591, 0.00923333, 0.00000092; 0.00001438, 0.00000092, 0.00302151],10000)

Could anybody also explain if the cholesky decomposition is built into the MVNRND function or must it be run before the MVNRND, and if so any assistance on how to go about it would be greatly appreciated.

Regards Desperate Masters Student