11.1. Heun’s Method

In scientific computing and engineering, many real-world problems are modeled using ordinary differential equations (ODEs). However, not all ODEs can be solved analytically. To address this, numerical methods such as Euler's method, Runge-Kutta methods, and Heun’s method (also known as the improved Euler's method or the explicit trapezoidal rule) are used. Heun's Method is a second-order numerical technique for solving initial value problems (IVPs) in ODEs. It provides better accuracy than Euler’s method by considering the average of the slopes at the beginning and predicted endpoint of an interval. It is particularly useful in practical engineering problems where stability and accuracy are essential.

Heun's Method is a numerical technique used to solve ordinary differential equations (ODEs) that can't be solved analytically. Unlike analytical methods, which give exact solutions, numerical methods provide approximate solutions by using computations. Heun’s Method is second-order, which means it has a higher accuracy compared to first-order methods like Euler’s method. This increased accuracy is achieved by taking the average of the slopes calculated at both the beginning and the end of each interval when predicting the next value.

Think of trying to predict the height of a plant that grows at varying rates. If you measure the height at the start and end of a day, the average height change gives a more accurate prediction than simply using the starting point to predict the next day’s height, which is what Euler's method does.

Consider a first-order initial value problem: 𝑑𝑦 / 𝑑𝑥 = 𝑓(𝑥,𝑦), 𝑦(𝑥0 )= 𝑦0. We aim to approximate the value of 𝑦 at discrete points using Heun’s method.

Heun’s Method is typically applied to first-order initial value problems, which can be modeled by the equation dy/dx = f(x, y). This represents a relationship where the rate of change of y with respect to x is determined by a function f. The goal is to find the approximations of y at various points starting from a known initial value, y0, at x0.

Imagine you're plotting the path of a car based on its speed. If you know the speed at the start of the hour, you can make a rough prediction of where it will be at the end, but by considering how the speed might change throughout the hour, you can get a much closer estimate of its final position.

Given the initial condition (𝑥0, 𝑦0), and a step size ℎ, Heun’s method computes the next value 𝑦𝑛+1 as follows: Step 1: Predictor (Euler's estimate) 𝑦∗ = 𝑦𝑛 + ℎ𝑓(𝑥𝑛,𝑦𝑛) Step 2: Corrector 𝑦𝑛+1 = 𝑦𝑛 + [𝑓(𝑥𝑛,𝑦𝑛)+ 𝑓(𝑥𝑛+1,𝑦∗)] * ℎ / 2. Here, 𝑦∗ is the predicted value using Euler’s method, and the corrector refines this prediction using the average of the slopes at the beginning and end of the interval.

Heun’s Method involves two main steps: First, you make a prediction of the next y value (called 𝑦∗) using Euler's method. This is your first estimate where you only consider the initial slope. Second, you correct this value by taking into account the slope at this predicted endpoint. The final value is determined by averaging these two slopes, which helps produce a more accurate estimate of the y value at that point.

Consider a teacher predicting a student's test score based on their performance in earlier tests. Initially, they guess based on the last test's score, much like the predictor step. Then, they revise that estimate by also considering how well the student performed on all previous tests together, leading to a more informed final prediction.

Let us solve over the interval [𝑥0,𝑥𝑛] with step size ℎ:

1. Initialize: Set 𝑥0,𝑦0,ℎ, and number of steps 𝑛.
2. Loop for each step:
   • Compute the predictor: 𝑦∗ = 𝑦𝑛 + ℎ ⋅𝑓(𝑥𝑛,𝑦𝑛)
   • Compute the corrector: 𝑦𝑛+1 = 𝑦𝑛 + ℎ [𝑓(𝑥𝑛,𝑦𝑛)+𝑓(𝑥𝑛 + ℎ,𝑦∗)] / 2
   • Update: 𝑥𝑛+1 = 𝑥𝑛 + ℎ.

This algorithm provides a structured way to apply Heun’s Method in any scenario. Start by defining your initial variables: the starting x and y values, the step size ℎ, and the number of steps to perform. For each step in the loop, first compute a prediction of the next value, then refine it through the corrector step, and finally update your variable x for the next iteration. This cycle continues until you reach the desired endpoint of your interval.

Think of this algorithm like baking a cake. You start with the basic ingredients (initial conditions), follow a step-by-step recipe (the loop), taste-test your batter (the predictor), adjust the flavors (the corrector), and then move on to the next layer once it’s done (updating x). This creates a sequence of improved results until the cake is fully baked.

Given: 𝑑𝑦 / 𝑑𝑥 = 𝑥 + 𝑦, 𝑦(0) = 1. Find: 𝑦 at 𝑥 = 0.1 using Heun’s Method with ℎ = 0.1. Step 1: 𝑓(𝑥0,𝑦0) = 𝑓(0,1) = 0 + 1 = 1. 𝑦∗ = 1 + 0.1 × 1 = 1.1. Step 2: 𝑓(𝑥1,𝑦∗) = 𝑓(0.1,1.1) = 0.1 + 1.1 = 1.2. 𝑦1 = 1 + 0.1 * (1 + 1.2) / 2 = 1 + 0.05 × 2.2 = 1.11. Result: 𝑦(0.1) ≈ 1.11.

In this example, we use Heun's Method to solve a specific ODE. First, we compute the function value at the initial condition. The predictor gives us a preliminary value for y. The corrector step then averages the slopes from the start to the end of the interval to compute the more accurate next value. After the calculations, we find that the approximate value of y at x = 0.1 is about 1.11.

It's like predicting travel time based on speed. You calculate how far light will travel in a specific time (predictor) but then consider the actual stops along the way and average the speeds (corrector) for a more realistic travel time.

Heun’s method can be viewed as applying the trapezoidal rule for numerical integration. Instead of using a single tangent (like Euler's method), it averages the slope at the beginning and the predicted endpoint of the interval. This generally reduces the error significantly.

Geometrically, Heun's Method can be imagined as drawing a trapezoid under the curve of the function f(x, y). Instead of estimating the area under the curve with just one side (the initial slope), Heun's Method uses the average of the slopes to create a better approximation, which leads to a smaller error compared to methods like Euler’s, where only one slope is considered.

Imagine you're trying to measure the amount of rainfall by looking at a curved path of water on a hillside. If you only estimate based on the height of the first spot (like in Euler's Method), your measurement will be off. However, if you estimate both the beginning and the end points and average them out (as in Heun’s Method), your overall measurement of total water flow would be much more accurate.

Feature | Euler’s Method | Heun’s Method

No detailed explanation available.

No real-life example available.