Mathematics Foundations/18.3 Numerical Integration and Differentiation

Numerical Integration and Differentiation Introduction Numerical integration and differentiation are fundamental techniques in numerical analysis that approximate the calculus operations when analytical solutions are difficult or impossible to obtain. These methods are essential in scientific computing, engineering simulations, and data analysis. Numerical Differentiation Finite Difference Methods Finite difference methods approximate derivatives using differences of function values at discrete points. Forward Difference The forward difference approximation of the first derivative is: where is a small step size. This approximation has an error of order . Backward Difference The backward difference approximation is: This also has an error of order . Central Difference The central difference approximation provides better accuracy: with an error of order . Higher Derivatives For the second derivative, we can use: Richardson Extrapolation Richardson extrapolation improves the accuracy of numerical differentiation by combining approximations with different step sizes: where is a finite difference approximation with step size . Numerical Integration Newton-Cotes Formulas Newton-Cotes formulas approximate definite integrals by fitting polynomials to function values at equally spaced points. Trapezoidal Rule The trapezoidal rule approximates the integral using linear interpolation: For subintervals: where and . The error is of order . Simpson's Rule Simpson's rule uses quadratic interpolation: For subintervals (where is even): The error is of order , making it generally more accurate than the trapezoidal rule. Simpson's 3/8 Rule For three subintervals of equal length: Gaussian Quadrature Gaussian quadrature methods select optimal evaluation points to achieve higher accuracy: where are the roots of the Legendre polynomials and are the corresponding weights. Adaptive Quadrature Adaptive quadrature methods adjust the step size based on the function behavior: Algorithm: 1. Estimate the integral over an interval 2. Divide the interval in half and estimate the integral over each half 3. If the sum of the two half-interval estimates is close enough to the whole-interval estimate, accept it 4. Otherwise, recursively apply the algorithm to each half-interval Romberg Integration Romberg integration combines the trapezoidal rule with Richardson extrapolation: Error Analysis Truncation Error Truncation error arises from approximating continuous operations with discrete methods: For the trapezoidal rule: For Simpson's rule: where is some point in . Round-off Error Round-off error occurs due to finite precision arithmetic and increases with the number of function evaluations. Applications Solving Differential Equations Numerical integration is essential for solving initial value problems: Numerical Quadrature in Physics Applications include calculating: Electric and magnetic fields Fluid dynamics Quantum mechanical expectation values Data Analysis Numerical differentiation is used for: Finding rates of change in experimental data Signal processing Image edge detection Exercises 1. Use the trapezoidal rule with n = 4 to approximate . Compare with the result using Simpson's rule. 2. Derive the error term for the central difference approximation of the first derivative. 3. Implement the adaptive quadrature algorithm to approximate with an error tolerance of 10^-6. 4. Using the forward, backward, and central difference methods, approximate the derivative of at with . Compare with the exact value. 5. Apply Richardson extrapolation to improve the trapezoidal rule approximation of . 6. Derive Simpson's rule by integrating a quadratic interpolating polynomial. 7. Compare the computational efficiency and accuracy of Gaussian quadrature versus the trapezoidal rule for . 8. Analyze how the error in the trapezoidal rule changes as you double the number of subintervals from n = 2 to n = 128 for .