Welcome to Problem Solving in Numerical Methods!
In your A Level journey, you’ve spent a lot of time finding "exact" answers. However, in the real world—whether you are an engineer designing a bridge or an economist predicting market trends—equations are often too messy to solve perfectly. This is where Numerical Methods come in.
In this chapter, we focus on Problem Solving. This means taking all the tools you’ve learned (like finding roots and approximating areas) and applying them to real-life scenarios. Don't worry if it feels a bit "approximate" at first; the goal here is to find an answer that is accurate enough to be useful!
1. Locating Roots in a Real-World Context
The first step in solving many real-world problems is finding where a function equals zero. We call this locating a root. In a context-based question, \(f(x) = 0\) might represent the time when a projectile hits the ground or the production level where a company stops losing money.
How to do it: The Change of Sign Method
If a function \(f(x)\) is continuous (it has no breaks or jumps) and you find two numbers, \(a\) and \(b\), such that \(f(a)\) is negative and \(f(b)\) is positive, there must be at least one root between them.
Example: A ball's height is given by \(h(t) = -4.9t^2 + 20t + 2\). To find when it hits the ground, we look for \(h(t) = 0\). If \(h(4) = 3.6\) and \(h(5) = -20.5\), we know the ball hits the ground between 4 and 5 seconds because the sign changed.
Quick Review: When this fails
- Touching the axis: If the graph just bounces off the x-axis, there is no change of sign, even though there is a root!
- Asymptotes: If the graph jumps from positive to negative across a vertical "break" (like \(y = 1/x\)), the sign changes, but there is no root there.
- Multiple roots: If there are two roots very close together, the sign might change and then change back, making you miss them entirely.
Key Takeaway: A change of sign proves a root exists only if the function is continuous in that interval.
2. Iteration: The Art of Getting Closer
Once you know a root exists in an interval, you need to "zoom in" on it. We do this using iteration, which is just repeating a mathematical process over and over to get a better answer.
Fixed Point Iteration (\(x = g(x)\))
To use this, we rearrange our original equation \(f(x) = 0\) into the form \(x = g(x)\). We then use a starting value (\(x_0\)) and plug it into the formula repeatedly: \(x_{n+1} = g(x_n)\).
The "Ladder" Analogy: Imagine trying to reach a high window. Each step of iteration is like climbing one rung higher on a ladder. If your formula is good, each step gets you closer to the target (the root).
Newton-Raphson Method
This is a more powerful way to find roots using tangents. The formula is:
\(x_{n+1} = x_n - \frac{f(x_n)}{f'(x_n)}\)
Step-by-Step Process:
1. Pick a starting value \(x_0\) close to the root.
2. Calculate \(f(x_0)\) and the gradient \(f'(x_0)\).
3. Use the formula to find \(x_1\).
4. Use the ANS key on your calculator to repeat the process until the numbers stop changing!
Memory Trick: Newton-Raphson is like "sliding down the tangent." You start at a point on the curve and follow the tangent line down to the x-axis to find your next, better guess.
Key Takeaway: Newton-Raphson is usually faster but can fail if your starting point is near a stationary point (where the gradient is zero), as you'll be "shot off" to infinity!
3. Approximating Areas: The Trapezium Rule
In some problems, you need to find the area under a curve (integration), but the function is too hard to integrate normally. We use the Trapezium Rule to estimate it by splitting the area into strips that look like trapeziums.
Over-estimates and Under-estimates
MEI examiners love to ask if your answer is too high or too low. This depends on the curvature (concavity) of the graph:
- If the curve is concave upwards (like a valley \(\cup\)), the trapezium tops sit above the curve. This is an over-estimate.
- If the curve is concave downwards (like a hill \(\cap\)), the trapezium tops sit below the curve. This is an under-estimate.
Using Rectangles
You can also use simple rectangles to find bounds:
1. Lower Bound: Use the height of the curve at the "lower" side of each strip.
2. Upper Bound: Use the height of the curve at the "higher" side of each strip.
Quick Review Box:
More strips = Better accuracy. If a problem asks you to improve an estimate, the easiest answer is usually to "increase the number of strips (n)."
Key Takeaway: The true area always lies between the lower bound (rectangles) and the upper bound (rectangles), and the Trapezium Rule usually sits somewhere in the middle.
4. Solving Problems in Context (Me6)
When a question gives you a "story" (e.g., a population model or a chemical reaction rate), follow these steps to stay calm:
Step 1: Translate the words.
If the question says "Find when the population reaches 5000," and the model is \(P(t)\), you are solving \(P(t) = 5000\). To use numerical methods, rewrite it as \(f(t) = P(t) - 5000 = 0\).
Step 2: Check the units.
Are you working in years? Thousands of people? Radians? (Always check your calculator is in Radians for calculus-based numerical methods!)
Step 3: Evaluate the model.
Numerical methods give you a number, but does that number make sense? If you get a negative time for a ball's flight, you've found a mathematical root that isn't a physical reality.
Did you know? Modern GPS systems use iteration to calculate your position. Your phone doesn't solve the equations once; it repeats a numerical process thousands of times a second to "zoom in" on where you are on the map!
Key Takeaway: Always relate your final numerical answer back to the original units and context of the question.
Final Summary for Revision
- Locating Roots: Use change of sign; watch out for discontinuities or vertical asymptotes.
- Iteration: \(x = g(x)\) produces cobweb or staircase diagrams. Newton-Raphson uses tangents.
- Integration: Use the Trapezium Rule. Check the shape (\(\cup\) vs \(\cap\)) to decide if it's an over or under-estimate.
- Real Life: Always keep an eye on your units and ensure your answer is "sensible" for the situation described.
Don't worry if these methods feel a bit repetitive—that's exactly what they are designed to be! With practice, you'll start to see the patterns. Good luck!