Mathematics of Computation · 1964 · 74 citations · 8 references
ComputingComputer ScienceComputational GeometryComputational PhysicsScientific ComputingIntegral Equation FormulationsIntegral EquationsNumerical AlgorithmsNumerical AnalysisNumerical IntegrationNumerical Methods For Differential EquationsHigh-dimensional IntegrationHigh-order Numerical IntegrationNumerical Integration FundamentalsNumerical Methods For Integral EquationsQuadrature RulesSimpson's RuleTrapezoidal Approximation
the first column gives the number of points used in the calculation. The results in the last column, although not as good as Simpson, are considerably better than those obtained by the plain trapezoid rule. Reason for the improvement is obvious: the integrand of (2) has the property (shared with the Jacobian of transformation (ii)) that it vanishes, together with all its derivatives, at the endpoints of the interval of integration. As we shall frequently encounter this property, we shall briefly refer to it as property P. For integrands with property P the trapezoid rule is just as good as Simpson. For if f(x) has property P then we can replace f f(x) dx by either of the integrals 1 ~ ~~~~~~~ 1?1/ N (3) I f(x) dx, f*(x) dx
8
A method for numerical integration
C. B. Haselgrove · Mathematics of Computation · 1961 · 206 citations · Full text