Numerical Methods for the FE Exam
Numerical methods provide approximate solutions to mathematical problems that are too complex or impossible to solve analytically. The FE exam tests a focused set of numerical techniques that every engineer should know. These methods are practical tools that bridge the gap between theory and real-world problem solving.
Root Finding: Solving f(x) = 0
Root-finding methods are used to find the values of x that make a function equal to zero. The FE exam may test these methods:
Bisection Method
The bisection method is the simplest root-finding technique:
- Start with two points a and b where f(a) and f(b) have opposite signs (meaning a root exists between them by the intermediate value theorem)
- Evaluate f at the midpoint c = (a + b) / 2
- If f(c) has the same sign as f(a), replace a with c; otherwise replace b with c
- Repeat until the interval is sufficiently small
The bisection method always converges, but it converges slowly. Each iteration cuts the interval in half.
Newton-Raphson Method
Newton-Raphson is faster than bisection but requires knowing the derivative:
x_(n+1) = x_n - f(x_n) / f'(x_n)
- Start with an initial guess x_0
- Apply the formula iteratively until convergence
- Converges quickly (quadratically) when the initial guess is close to the root
- May diverge or oscillate if the initial guess is poor or if f'(x) is near zero
On the FE exam, you may need to perform one or two iterations of Newton-Raphson to find an approximate root.
Secant Method
Similar to Newton-Raphson but does not require the derivative:
x_(n+1) = x_n - f(x_n) * (x_n - x_(n-1)) / (f(x_n) - f(x_(n-1)))
Uses two previous points to approximate the derivative. Convergence is faster than bisection but slower than Newton-Raphson.
Numerical Integration
When you cannot find the antiderivative of a function, numerical integration provides an approximate value of the definite integral.
Trapezoidal Rule
Approximates the area under the curve using trapezoids. For n equally spaced subintervals with width h = (b - a) / n:
Integral approximately equals (h/2) * [f(x_0) + 2*f(x_1) + 2*f(x_2) + ... + 2*f(x_(n-1)) + f(x_n)]
The error decreases as you increase the number of subintervals. The trapezoidal rule is exact for linear functions.
Simpson's Rule
Uses parabolic segments instead of straight lines for better accuracy. Requires an even number of subintervals:
Integral approximately equals (h/3) * [f(x_0) + 4*f(x_1) + 2*f(x_2) + 4*f(x_3) + ... + 4*f(x_(n-1)) + f(x_n)]
Notice the alternating pattern of coefficients: 1, 4, 2, 4, 2, ..., 4, 1. Simpson's rule is exact for polynomials up to degree 3.
Interpolation
Interpolation estimates values between known data points.
Linear Interpolation
The simplest method: connect two known points with a straight line and find the value at the desired x:
y = y_1 + (x - x_1) * (y_2 - y_1) / (x_2 - x_1)
This is frequently used on the FE exam when reading values from tables that do not have the exact entry you need.
Lagrange Interpolation
A general method that passes a polynomial through n data points. For two points it reduces to linear interpolation. For three points, it gives quadratic interpolation. The formula is in the reference handbook.
Numerical Differentiation
Approximate derivatives from discrete data points:
- Forward difference: f'(x) approximately equals [f(x + h) - f(x)] / h
- Backward difference: f'(x) approximately equals [f(x) - f(x - h)] / h
- Central difference: f'(x) approximately equals [f(x + h) - f(x - h)] / (2*h). More accurate than forward or backward differences.
Solving ODEs Numerically: Euler's Method
Euler's method provides a simple numerical solution to first-order ODEs of the form dy/dx = f(x, y):
y_(n+1) = y_n + h * f(x_n, y_n)
Where h is the step size. Starting from an initial condition (x_0, y_0), the method steps forward iteratively to approximate the solution. Euler's method is first-order accurate, meaning smaller step sizes give better results but require more computation.
Runge-Kutta Methods
The fourth-order Runge-Kutta method (RK4) is more accurate than Euler's method and is the most commonly used numerical ODE solver. It uses four evaluations of f(x, y) per step to achieve fourth-order accuracy. The formulas are in the reference handbook; on the FE exam, you are more likely to use Euler's method due to its simplicity.
Error Analysis
Numerical methods introduce errors. Key concepts:
- Truncation error: Error from approximating a mathematical procedure (using finite terms of a series, using finite step sizes).
- Round-off error: Error from representing real numbers with finite precision.
- Convergence: The property that the approximate solution approaches the exact solution as the step size decreases.
- Stability: The property that errors do not grow without bound during computation.
When to Use Each Method
- Root finding: When you need to solve f(x) = 0 and cannot solve algebraically
- Numerical integration: When you cannot find an antiderivative, or when the function is defined only by data points
- Interpolation: When you need a value between entries in a table
- Numerical ODE solving: When a differential equation cannot be solved analytically
Numerical methods questions on the FE exam are typically straightforward: apply the formula, perform a few iterations, and find the approximate answer. Practice the mechanical steps of each method so you can execute them quickly on exam day.