function result = p00_monte_carlo ( problem, order ) %*****************************************************************************80 % %% P00_MONTE_CARLO applies a Monte Carlo procedure to Hermite integrals. % % Discussion: % % We wish to estimate the integral: % % I(f) = integral ( -oo < x < +oo ) f(x) exp ( - x * x ) dx % % We do this by a Monte Carlo sampling procedure, in which % we select N points X(1:N) from a standard normal distribution, % and estimate % % Q(f) = sum ( 1 <= I <= N ) f(x(i)) / sqrt ( pi ) % % Licensing: % % This code is distributed under the GNU LGPL license. % % Modified: % % 17 May 2010 % % Author: % % John Burkardt % % Parameters: % % Input, integer PROBLEM, the index of the problem. % % Input, integer ORDER, the order of the Gauss-Laguerre rule % to apply. % % Output, real RESULT, the approximate integral. % seed = 123456789; [ x_vec, seed ] = r8vec_normal_01 ( order, seed ); option = 2; f_vec = p00_fun ( problem, option, order, x_vec ); weight = order / sqrt ( 2.0 * pi ); result = sum ( f_vec(1:order) ) / weight; return end