Introduction: Why Numerical Methods?

Welcome to the world of Numerical Methods! In your previous maths studies, you’ve mostly looked for "exact" solutions—like finding that \(x = 2\). However, in the real world (and in Further Maths), some differential equations are so complex that they are impossible to solve exactly using standard algebra.

That is where numerical methods come in. Think of these as "smart guessing" techniques. Instead of finding a perfect formula for a curve, we calculate specific points along that curve step-by-step. It’s like using a series of short, straight lines to map out a winding mountain path. In this chapter, we will learn how to estimate solutions for both first-order and second-order differential equations.

1. The Building Blocks: What is \(h\)?

Before we look at the formulas, we need to understand the step size, usually called \(h\).

Imagine you are walking along a graph. If you take huge leaps (\(h\) is large), you’ll finish quickly but might miss the shape of the curve. If you take tiny baby steps (\(h\) is small), your path will be much closer to the real curve, but it will take a lot more calculations!

Quick Terms:
• \(x_n\): Your current position on the x-axis.
• \(x_{n+1}\): Your next position (\(x_n + h\)).
• \(y_n\): The value of the solution at \(x_n\).
• \(y_{n+1}\): The value we want to find!

Memory Aid: Just remember that the subscript \(n\) means "now" and \(n+1\) means "next".

2. Solving First-Order Differential Equations

A first-order equation involves the first derivative, \( \frac{dy}{dx} \). We have two main ways to approximate this gradient.

Method A: The Forward Difference

This is the simplest method. It uses the gradient at your current point to predict the next point. It’s essentially the same as the "rise over run" formula you learned in GCSE!

The Formula:
\( (\frac{dy}{dx})_n \approx \frac{y_{n+1} - y_n}{h} \)

How to use it:
Usually, you are given an equation for \( \frac{dy}{dx} \). You rearrange the formula above to find \(y_{n+1}\):
\( y_{n+1} \approx y_n + h(\frac{dy}{dx})_n \)

Method B: The Central Difference

The central difference is often more accurate because it looks at the points behind and ahead of where you are to find a balanced gradient.

The Formula:
\( (\frac{dy}{dx})_n \approx \frac{y_{n+1} - y_{n-1}}{2h} \)

Don’t worry if this seems tricky! The "2h" on the bottom is there because the distance from \(y_{n-1}\) to \(y_{n+1}\) is two steps of \(h\).

Key Takeaway: Use Forward Difference if you only know the current point. Use Central Difference for better accuracy if you have information about the previous point too.

3. Solving Second-Order Differential Equations

When an equation involves \( \frac{d^2y}{dx^2} \), we need an approximation for the second derivative. This helps us model things like acceleration or the curvature of a bridge.

The Formula:
\( (\frac{d^2y}{dx^2})_n \approx \frac{y_{n+1} - 2y_n + y_{n-1}}{h^2} \)

Step-by-Step Process:
1. Replace the \( \frac{d^2y}{dx^2} \) in your differential equation with the approximation above.
2. Replace any \( \frac{dy}{dx} \) terms with the Central Difference approximation.
3. Replace \(x\) and \(y\) with \(x_n\) and \(y_n\).
4. Rearrange the whole mess to get \(y_{n+1}\) on its own.

Did you know?
These methods are the ancestors of the algorithms used today in weather forecasting and video game physics engines to calculate how objects move over time!

4. Avoiding Common Mistakes

Even the best Further Mathematicians can trip up here. Watch out for these:

  • The 2h Trap: In the Central Difference formula for \( \frac{dy}{dx} \), students often forget to divide by \(2h\) and divide by \(h\) instead. Don't let that be you!
  • Calculator Radians: If your differential equation contains trig functions (like \( \sin(x) \)), make sure your calculator is in Radians.
  • Rounding too early: Numerical methods involve many steps. If you round your numbers at each step, your final answer will be "polluted" by rounding errors. Keep as many decimals as possible until the very end!

Quick Review Box

Forward Difference: \( (\frac{dy}{dx})_n \approx \frac{y_{n+1} - y_n}{h} \)
Central Difference: \( (\frac{dy}{dx})_n \approx \frac{y_{n+1} - y_{n-1}}{2h} \)
Second Derivative: \( (\frac{d^2y}{dx^2})_n \approx \frac{y_{n+1} - 2y_n + y_{n-1}}{h^2} \)

Final Tip: When you get an exam question, usually the first step is just substituting these approximations into the differential equation given. Do that first to bank some easy marks!