This discussion focuses on the decomposition of functions for numerical integration, specifically using Fourier and Legendre decomposition in conjunction with the Gauss Quadrature method. Participants emphasize the importance of adapting the number of quadrature points based on the function's behavior, suggesting a trial and error approach to achieve stable integration results. The conversation highlights that more quadrature points should be allocated to regions with complex variations, while fewer points can be used in smoother regions.
PREREQUISITESMathematicians, physicists, and engineers involved in numerical analysis and integration techniques, particularly those utilizing Gauss Quadrature for complex functions.
ecastro said:Is their a tutorial or a reference on how to decompose a function, specifically Fourier and Legendre decomposition, for numerical integration? The method I am going to use for the numerical integration is the Gauss Quadrature, and I suppose I need to decompose my function for the rule to work.