24 January 2008 12:00:53 PM INT_EXACTNESS_GEN_HERMITE C++ version Investigate the polynomial exactness of a generalized Gauss-Hermite quadrature rule by integrating exponentially weighted monomials up to a given degree over the (-oo,+oo) interval. INT_EXACTNESS_GEN_HERMITE: User input: Quadrature rule X file = "gen_herm_o8_a1.0_modified_x.txt". Quadrature rule W file = "gen_herm_o8_a1.0_modified_w.txt". Quadrature rule R file = "gen_herm_o8_a1.0_modified_r.txt". Maximum degree to check = 18 Power of |X|, ALPHA = 1 OPTION = 1, integrate f(x) Spatial dimension = 1 Number of points = 8 The quadrature rule to be tested is a generalized Gauss-Hermite rule ORDER = 8 ALPHA = 1 OPTION = 1: Modified rule: Integral ( -oo < x < +oo ) f(x) dx is to be approximated by sum ( 1 <= I <= ORDER ) w(i) * f(x(i)). Weights W: w[ 0] = 1.058214197948879 w[ 1] = 0.8524080381127395 w[ 2] = 0.7750492008314336 w[ 3] = 0.7331317124710707 w[ 4] = 0.7331317124710707 w[ 5] = 0.7750492008314336 w[ 6] = 0.8524080381127395 w[ 7] = 1.058214197948879 Abscissas X: x[ 0] = -3.065137992375079 x[ 1] = -2.129934340988268 x[ 2] = -1.321272530993643 x[ 3] = -0.5679328213965031 x[ 4] = 0.5679328213965031 x[ 5] = 1.321272530993643 x[ 6] = 2.129934340988268 x[ 7] = 3.065137992375079 Region R: r[ 0] = -1e+30 r[ 1] = 1e+30 A generalized Gauss-Hermite rule would be able to exactly integrate monomials up to and including degree = 15 Error Degree 4.440892098500626e-16 0 2.547875105340935e-17 1 4.440892098500626e-16 2 6.852157730108388e-17 3 6.661338147750939e-16 4 8.326672684688674e-17 5 1.036208156316813e-15 6 2.220446049250313e-16 7 8.881784197001252e-16 8 8.881784197001252e-16 9 9.473903143468002e-16 10 0 11 7.894919286223335e-16 12 0 13 7.218211918832764e-16 14 0 15 0.01428571428571392 16 0 17 0.06507936507936508 18 INT_EXACTNESS_GEN_HERMITE: Normal end of execution. 24 January 2008 12:00:53 PM