Numerical Techniques Exam Paper
Uploaded by
Prathamesh HendreNumerical Techniques Exam Paper
Uploaded by
Prathamesh HendreCommon questions
Powered by AIUsing Euler's Method with step size h = 0.1 for dy/dx = 1 + x and y(0) = 1, the iterative formula is y_{n+1} = y_n + h(f(x_n, y_n)). Starting with y(0) = 1, y(0.1) = y(0) + 0.1(1 + 0) = 1 + 0.1 = 1.1. Therefore, y(0.1) is approximately 1.1 .
To evaluate the integral ∫(1/(x^2+1)) dx from 0 to 1 using the Trapezoidal Rule with step size h = 0.5, we subdivide the interval into subintervals [0, 0.5] and [0.5, 1]. The trapezoidal rule calculates the area as approximately (h/2)[f(0) + 2f(0.5) + f(1)]. Substituting the function values, we get (0.5/2)[1 + 2*(1/1.25) + 1/2] = 0.643. Thus, the approximated integral value is 0.643 .
In the Newton-Raphson method, the iterative formula used is x_{n+1} = x_n - f(x_n)/f'(x_n). Using f(x) = x^2 - 12 and starting with x0 = 3.4, we first calculate f'(x) = 2x. Thus, x1 = 3.4 - ((3.4^2 - 12)/(2 * 3.4)) = 3.4647. Repeating this process, x2 ≈ 3.4641. Therefore, through two iterations, the approximation of √12 becomes approximately 3.4641 .
The False Position Method often converges slower because it uses a linear approximation based on the secant line between two points, which may not quickly adjust to the curvature of the actual function. For example, with x^3 - x^2 - 2 = 0, we start with two initial guesses, say x=1 and x=2. The intersection of the secant line with the x-axis provides a new estimate for the root. This process may not efficiently move towards the true root due to the function's non-linear nature, requiring more iterations than methods like Newton-Raphson, which utilize the derivative for faster convergence .
A divided difference table is constructed by establishing a column for divided differences for each order. For given points (x : 2, 4, 9, 10; y: 4, 56, 711, 980), the first column contains the y-values. The second column contains the first divided differences, calculated as (f[x1,x2] - f[x0,x1])/(x2 - x0), and so forth for higher orders. This table facilitates interpolation, as the polynomial is expressed using the divided differences and the corresponding x-values .
Simpson's 3/8 Rule is derived by integrating a cubic function p(x) that passes through three equally spaced points using polynomial interpolation. The rule approximates the integral of a function as (3h/8)[f(x0) + 3f(x1) + 3f(x2) + f(x3)], where h is the interval width. It is accurate for polynomial integrands up to cubic degree because it exactly integrates cubic polynomials due to its derivation from such functions. This higher degree of exactitude improves accuracy for functions well-approximated by a cubic polynomial between the chosen points .
Newton's Backward Difference Interpolation Formula is applicable for interpolating the value of a function at a specific point using backward differences. It is expressed as f(x) = f(x_n) + ∇f(x_n)(x - x_n)/1! + ∇^2f(x_n)(x - x_n)(x - x_{n-1})/2! + ... In this formula, ∇ represents the backward difference operator and is used when data points are equally spaced. This method is particularly effective when calculating the values closer to the end of the data set .
Euler's Modified Method (also known as the Heun method) improves accuracy by correcting the slope estimate used in the standard Euler's Method. It calculates an initial prediction using Euler's explicit formula and then refines this using the average of slopes at the start and end of the interval. This results in a two-step predictor-corrector approach, reducing the local truncation error significantly compared to standard Euler's Method, particularly for equations with rapidly changing dynamics .
Relative error is the absolute error divided by the true value, providing a measure of error in relation to the size of the exact value. Percentage error is the relative error multiplied by 100, giving the error as a percentage. This conversion allows errors to be compared across different scales and magnitudes .
In numerical analysis, the forward difference operator Δ and the shift operator E are related by the equation Δ = E - 1. The operator Δ indicates the difference between successive terms, i.e., Δy_n = y_{n+1} - y_n, whereas E represents the operation of shifting a function by one unit, i.e., Ey_n = y_{n+1}. Therefore, applying E to y_n and subtracting 1 times y_n results in the forward difference, thus Δ = (E - 1)y_n .