Slope fields, Euler’s method and algorithms: VCE Specialist Mathematics Units 3 & 4

1. Read a differential equation as a map of local slopes

For a first-order differential equation y′ = F(x, y), every allowed point (x, y) is assigned a slope. A slope field displays short line segments with those slopes on a grid. The segments describe local direction, not function values and not finite jumps. A solution curve should pass through them tangentially while also passing through its specified initial point.

At a point where F(x, y) = 0, the segment is horizontal. A positive value gives an upward-sloping segment from left to right, while a negative value gives a downward-sloping segment. A large magnitude gives a steep segment. Fields are usually drawn with equal segment lengths, so the visual length does not encode the rate's magnitude.

Test a few informative points when matching an equation to a field. If F depends only on x, slopes are the same along each vertical grid line. If F depends only on y, slopes are the same along each horizontal grid line. If F depends on both variables, neither pattern is generally present. These structural observations often eliminate incorrect choices faster than inspecting every segment.

2. Equilibria and zero slopes are not the same claim

An equilibrium solution is a constant function y = c that satisfies the differential equation throughout the relevant x-interval. It requires F(x, c) = 0 for every x in that interval. A single horizontal segment at one point is insufficient. For y′ = x − y, the slopes vanish on y = x, but that diagonal is not a solution: its own derivative is 1, while the equation would require zero.

A zero-slope curve can still be informative. It marks where a solution may change from rising to falling as it passes through the plane. To investigate an autonomous equation y′ = g(y), mark the zeros of g and inspect its sign between them. The resulting direction pattern predicts whether nearby solutions move toward or away from an equilibrium level.

When a field omits points because F is undefined there, do not interpolate a solution straight across the missing region. The equation's domain comes before graphical guesses. Likewise, a small viewing window can make a rapidly growing solution appear to terminate even though it has simply left the display. Separate mathematical boundaries from plotting boundaries.

3. Euler's method repeatedly follows a tangent

Suppose the current approximation is (x_n, y_n) and the positive step size is h. The slope there is F(x_n, y_n). Following the tangent for a horizontal distance h gives the first-order Euler update:

The slope must be evaluated at the old point. Updating x first and then evaluating F with the new x but old y gives a different procedure. Compute the slope before changing either coordinate, or store both new coordinates separately. This detail matters especially when F depends explicitly on x.

The initial point is exact data, not the result of an Euler step. If the endpoint is b and (b − x₀)/h is an integer N, perform N updates. A table containing N + 1 points includes the initial point and the N computed points. Confusing point count with step count is a frequent off-by-one error.

Each update replaces the unknown short curve segment with a straight tangent step. The next slope is evaluated at the new approximation, which may already differ from the exact solution. Consequently errors can accumulate, and an exact rate formula does not make the resulting coordinates exact. Use the approximation symbol when reporting a numerical Euler estimate.

4. Step size affects accuracy, workload and behaviour

A smaller positive step usually improves an Euler approximation over a fixed interval when the solution and rate function behave suitably and numerical errors are controlled. It also requires more updates. Halving h doubles the number of steps to the same endpoint. Compare estimates at the same input value; estimates at different endpoints do not isolate the effect of step size.

Concavity explains the direction of a single tangent error. For a smooth concave-up exact solution, the tangent at an exact starting point lies below the nearby curve; for a concave-down solution it lies above. Across many steps, however, the starting point is approximate and the slope depends on that approximation. Do not claim a universal overestimate or underestimate solely from one tangent picture without checking the equation's behaviour.

A large step can produce implausible results even when the true solution is well behaved. For y′ = −2y, Euler gives y_n+1 = (1 − 2h)y_n. With h = 0.75 and positive y₀, the factor is −1/2 and the iterates alternate in sign, although the exact solution y₀e^−2x stays positive. The numerical method has created the sign changes.

Keep sufficient intermediate precision. Rounding every slope or updated value to two decimal places introduces additional errors that may hide the effect of changing h. Carry the stored calculator values and round the final requested estimate. If a task explicitly prescribes rounding at each step, follow that instruction and state that convention.

5. Design an algorithm that preserves the mathematics

An algorithm should specify its inputs, starting values, repeated calculation and output. For Euler's method, the essential inputs are F, x₀, y₀, h and the number of steps N. A clear pseudocode version stores the slope first and outputs the initial point before starting the loop:

INPUT x0, y0, h, N
x ← x0
y ← y0
OUTPUT x, y
FOR n FROM 1 TO N
    slope ← F(x, y)
    next_y ← y + h × slope
    x ← x + h
    y ← next_y
    OUTPUT x, y
END FOR

Trace one or two iterations by hand before trusting an implementation. For the equation x + y from worked example 2, the first output after the update must be (0.2, 1.2). If it is (0.2, 1.24), the program probably used the updated x in the slope. This is a targeted test that identifies the likely error rather than merely observing that the final answer differs.

For an endpoint that is not an exact multiple of h away, state how the final step is handled. One option is a shorter final step Δx = b − x. Do not overshoot b and label the resulting value y(b). If a supplied algorithm uses a specific stopping condition, trace that condition literally before deciding which point it returns.

6. Explain what a numerical result establishes

Euler estimates approximate a solution from initial data; they do not find roots of an equation. Newton iteration has a different purpose and update rule. A slope field visualises directions; it does not supply an exact function simply because a smooth curve can be drawn through it. Keep the representation and the task distinct.

A non-routine investigation can compare a mixing model's exact solution with Euler values at several step sizes. Record the final input, approximation, exact value when known and signed error. Ask whether the approximations preserve nonnegative amounts and stay within expected equilibrium bounds. An unexpected overshoot can reveal a coarse step or an implementation mistake before it becomes a contextual claim.

When technology is permitted, present enough of the update process to make the algorithm visible. Give h, the initial point, the requested endpoint and a table or clearly labelled sequence. Then interpret the final number with the correct units and an approximation sign. A bare calculator output does not communicate which numerical method was used.