Problem 13 - Advanced Engineering Mathematics

Question

Consider the boundary-value problem $$ y^{\prime \prime}+x y=0, \quad y^{\prime}(0)=1, \quad y(1)=-1 $$ (a) Find the difference equation corresponding to the differential equation. Show that for \(i=0,1,2, \ldots, n-1\), the difference equation yields \(n\) equations in \(n+1\) unknowns \(y_{-1}, y_{0}, y_{1}, y_{2}, \ldots, y_{n-1}\). Here \(y_{-1}\) and \(y_{0}\) are unknowns since \(y_{-1}\) represents an approximation to \(y\) at the exterior point \(x=-h\) and \(y_{0}\) is not specified at \(x=0\). (b) Use the central difference approximation (5) to show that \(y_{1}-y_{-1}=2 h\). Use this equation to eliminate \(y_{-1}\) from the system in part (a). (c) Use \(n=5\) and the system of equations found in parts (a) and (b) to approximate the solution of the original boundary-value problem.

Step-by-Step Solution

Answer

The solution involves discretizing the differential equation, forming a system of equations in terms of finite differences, eliminating extra unknowns, and solving the system for \(n=5\).

Step 1: Discretization of the Differential Equation

The differential equation is \(y'' + xy = 0\) with boundary conditions \(y'(0) = 1\) and \(y(1) = -1\). We use the central difference approximation for second derivatives: \(y'' \approx \frac{y_{i+1} - 2y_i + y_{i-1}}{h^2}\). For the term \(x_i y_i\), we use \(x_i = i \, h\). The difference equation becomes \(\frac{y_{i+1} - 2y_i + y_{i-1}}{h^2} + i \, h \, y_i = 0\). Rearranging gives: \[y_{i+1} = 2y_i - y_{i-1} - h^2 (i \, h \, y_i)\].

Step 2: Formulate the System of Equations

The difference equation needs to hold for each \(i = 0, 1, 2, \ldots, n-1\). This yields \(n\) equations: 1. For \(i=0\): \(y_1 = 2y_0 - y_{-1} - h^2 \cdot 0 \cdot y_0\)2. For \(i=1\): \(y_2 = 2y_1 - y_0 - h^2 \cdot 1h \cdot y_1\) (so forth up to \(i=n-1\)). These equations involve \(n+1\) unknowns: \(y_{-1}, y_0, y_1, \ldots, y_{n-1}\). Since there is no boundary condition for \(y(0)\), \(y_0\) is unknown. \(y_{-1}\) is introduced to approximate conditions at \(x = -h\).

Step 3: Eliminate \(y_{-1}\) using Central Difference for \(y'(0)\)

The derivative \(y'(0) = 1\) is approximated by the central difference: \(y'(0) \approx \frac{y_1 - y_{-1}}{2h}\). Using the boundary condition, this gives us \(\frac{y_1 - y_{-1}}{2h} = 1\), leading to \(y_1 - y_{-1} = 2h\). This equation allows us to express \(y_{-1}\) in terms of \(y_1\): \(y_{-1} = y_1 - 2h\). We substitute this into the system from Step 2 to eliminate \(y_{-1}\).

Step 4: Solve the System with \(n=5\)

With \(n=5\), the grid points can be \(x_0 = 0, x_1 = h, \ldots, x_5 = 1\) (where \(h = 0.2\)). We have the modified system from Steps 2 and 3:1. \(y_1 = 2y_0 - (y_1 - 2h)\)2. Continue with similar approach for \(y_2, y_3, \ldots, y_5\).3. Add the boundary condition at \(x=1\), which is \(y_5 = -1\).Solving this system using substitution and elimination methods, possibly using numeric tools since it involves up to degree five polynomials.

Key Concepts

Difference Equations

Difference equations are discrete analogs of differential equations. Instead of dealing with continuous variables, difference equations deal with variables at discrete points, often on a grid. This concept is crucial when working with boundary-value problems where solutions are investigated at specific intervals.

In the given exercise, the differential equation \( y'' + xy = 0 \) is transformed into a difference equation by discretizing the x-value positions. Applying a difference equation removes the need for solving the differential equation analytically over a continuous range. Instead, we estimate the values of \( y \) at specified points marked on a grid or lattice, making it suitable for numerical solution.

The conversion results in a discrete system that is easier to manipulate, especially when computing the solution via numerical methods. Transforming a differential equation into a difference equation also helps handle problems that aren’t easily solvable by symbolic computation, allowing for practical computational approaches such as iterative solving methods.

Central Difference Approximation

Central difference approximation is a technique used to estimate the derivatives of a function. This method is particularly useful in converting differential equations into difference equations, as seen in the boundary-value problem.

For a second derivative, the central difference formula is given by: \[ y'' \approx \frac{y_{i+1} - 2y_i + y_{i-1}}{h^2} \] where \( h \) is the step size between consecutive grid points. This calculation provides an average rate of change around a central point, making it more accurate than either forward or backward difference approximations.

In the exercise problem, central difference approximation allows the expression of the second derivative in a manner that fits naturally into a difference equation framework. It provides a way to handle the boundary condition at \( x=0 \), which doesn't have a specified \( y \) value. Instead, it's addressed through the derivative condition using the central difference itself, which leads to the relation \( y_1 - y_{-1} = 2h \). This relationship is crucial for further simplifying and solving the system of equations.

Discretization of Differential Equations

Discretization is the process of transforming a continuous mathematical model, such as a differential equation, into a discrete one. In mathematical computations, this conversion is essential for creating solvable systems via numerical analysis.

In the boundary-value problem, discretization involves converting the continuous domain of \( x \) into a series of points \( x_0, x_1, ..., x_n \). Each of these points represents a grid location where the function \( y \) is estimated. By replacing the differential equation with difference approximations (as discussed with the central difference method), the differential term \( y'' \) becomes approachable in terms of simple algebraic equations, allowing for numerical solving methods.

Discretization of differential equations enables the study and solution of complex boundary-value problems when analytical solutions are either impossible or impractical to deduce. By utilizing this method, we can simulate real-world systems finely, ensuring we achieve a useful approximation of the desired solution. This approach greatly benefits engineering and physics problems where boundary conditions influence solutions significantly, much like in the given exercise.