This is the simplest way. Best, Matteo #include <oxstd.oxh> main() { decl cn = 10000; // sample size decl vMu = <1;2;3>; // mean vector decl mSig = <5, 3, 1; 3, 6, 2; 1, 2, 7>; // covariance matrix decl mZ = rann(cn, rows(mSig)); // generate n x k matrix of standard normal variates decl mX = mZ*choleski(mSig)'+vMu'; // factorise covariance matrix by choleski and multiply standard normals, then sum means println("Actual vs. estimated mean vector", vMu', meanc(mX)); println("Actual vs. estimated covariance matrix", mSig, variance(mX)); } Il 13/07/2015 23:23, Natalia Manduca Ferreira ha scritto: > > Dear collegues, > > I need to generating multivariate normal random variables, but I > don't know how. Could anyone help me? > > Very thank you! > > -- > Natalia Manduca -- Matteo Pelagatti Department of Economics, Management and Statistics Università degli Studi di Milano-Bicocca Tel +39 02 64485834