Infinite Integrals by Gaussian Quadrature
|
12-19-2022, 08:11 PM
Post: #4
|
|||
|
|||
RE: Infinite Integrals by Gaussian Quadrature
I figured out how the numbers are derived. The trick is *no* transformation!
Again, we assume a=0, then later put back a (if needed) Assume f is a quintic polynomial, we solved for integral exact solution. XCas> P(x) := horner([A,B,C,D,E,F],x) // quintic polynomial XCas> R := int(e^-x * P(x), x=0 .. inf) 120*A + 24*B + 6*C + 2*D + E+F XCas> M := w1*P(z1) + w2*P(z2) + w3*P(z3) - R If weights/points correctly picked, M=0, for any P coefficients. For simplicity, we use identity matrix for P coefficients. XCas> M := normal([ M(A=1,B=0,C=0,D=0,E=0,F=0), M(A=0,B=1,C=0,D=0,E=0,F=0), M(A=0,B=0,C=1,D=0,E=0,F=0), M(A=0,B=0,C=0,D=1,E=0,F=0), M(A=0,B=0,C=0,D=0,E=1,F=0), M(A=0,B=0,C=0,D=0,E=0,F=1) ]) XCas> transpose(M) \(\left(\begin{array}{c}-120+w_{1} z_{1}^{5}+w_{2} z_{2}^{5}+w_{3} z_{3}^{5}\\-24+w_{1} z_{1}^{4}+w_{2} z_{2}^{4}+w_{3} z_{3}^{4}\\-6+w_{1} z_{1}^{3}+w_{2} z_{2}^{3}+w_{3} z_{3}^{3}\\-2+w_{1} z_{1}^{2}+w_{2} z_{2}^{2}+w_{3} z_{3}^{2}\\-1+w_{1} z_{1}+w_{2} z_{2}+w_{3} z_{3}\\-1+w_{1}+w_{2}+w_{3}\end{array}\right)\) Last equation, we have sum of weights = 1, matching OP 6 equations, 6 unknown, using previous post numbers as guess. XCas> fsolve(M=0,[w1,z1, w2,z2, w3,z3] = [0.311,0.127, 0.654,1.00, 0.00833,7.87], [w1,z1, w2,z2, w3,z3]) [0.278517733569, 2.29428036028, 0.711093009929, 0.415774556783, 0.0103892565016, 6.28994508294] Weights/Points now match OP (in different orders, but that's OK) |
|||
« Next Oldest | Next Newest »
|
Messages In This Thread |
Infinite Integrals by Gaussian Quadrature - Eddie W. Shore - 11-23-2022, 01:55 PM
RE: Infinite Integrals by Gaussian Quadrature - Albert Chan - 12-14-2022, 08:21 PM
RE: Infinite Integrals by Gaussian Quadrature - Albert Chan - 12-19-2022, 02:22 PM
RE: Infinite Integrals by Gaussian Quadrature - Albert Chan - 12-19-2022 08:11 PM
|
User(s) browsing this thread: 1 Guest(s)