Foundations — scipy.integrate — odeint, solve_ivp (RK45, DOP853), quad

Before you can read the parent note you must own every symbol it uses without stumbling. Below, each symbol is built from nothing: plain words → the picture → why the topic needs it. Read top to bottom; each rung stands on the one below.

1. The function $f$ — a machine that turns a number into a number

Picture a curve drawn on paper. The horizontal axis is the input $x$; the height of the curve above that point is the output $f(x)$.

Why the topic needs it: everything in this module is about a function. quad wants the area under $f$; the ODE solvers are handed a function that reports a slope. If "function" is fuzzy, nothing downstream makes sense.

2. The interval $[a,b]$ and the symbols $a,b$

On the picture from figure 1, $a$ and $b$ are two vertical fence-posts on the horizontal axis. The strip of area we care about lives between them.

Why the topic needs it: quad(f, a, b) asks for the area of $f$ only between these two posts. For ODEs the analogous pair is $[t_0,t_f]$ — the start and end time.

3. $\Delta x$ — a small width, and the idea of "tiny piece"

Look at the amber rectangle in figure 2: its width is $\Delta x$ and its height is $f(x_i)$ at some sample point $x_i$. Its area is $f(x_i)\Delta x$.

Why the topic needs it: the whole trick of numerical integration is chopping a smooth thing into many pieces each of width $\Delta x$, and later letting that width shrink toward zero.

4. The sum symbol $\sum$ — "add up all these pieces"

So ${\sum }_{i}f(x_i)\Delta x$ literally reads: take rectangle 1's area, plus rectangle 2's area, plus …. That is a Riemann sum — figure 2 shows all the amber rectangles being counted.

Why the topic needs it: quad is literally a weighted sum ${\sum }_i{w}_{i}f(x_i)$ — see the next symbol.

5. The integral sign ${\int }_a^{b}f(x)dx$

${\int }_a^{b}f(x)dx={\lim }_{\Delta x\to 0}{\sum }_{i}f(x_i)\Delta x$

WHAT this line says: the integral is the limit of the rectangle-sum as the slivers get infinitely thin. WHY the limit: thinner rectangles hug the curve better, so the error vanishes. WHAT IT LOOKS LIKE: figure 2's jagged rectangle-tops smoothing into the true curve as they multiply.

Why the topic needs it: this exact symbol is what quad(f, a, b) computes. No integral sign → no quad.

6. The weights $w_i$ and sample points $x_i$

In a plain Riemann sum every weight is the same ($w_i=\Delta x$). Smart rules (Gaussian quadrature) pick unequal weights and clever sample spots so that few points give big accuracy. The parent's Simpson's rule $\frac{h}{3}[f(-h)+4f(0)+f(h)]$ has weights $\frac{h}{3},\frac{4h}{3},\frac{h}{3}$ — the middle sample counts four times as much.

Why the topic needs it: quad uses adaptive Gauss–Kronrod weights; understanding "weighted sum" is what makes that phrase meaningful.

7. The derivative $\frac{dy}{dt}$ — a slope, a speed

The red line in figure 3 is tangent to the curve — it just grazes it. Its steepness IS $\frac{dy}{dt}$ at that point. Steep line = changing fast; flat line = barely changing.

Why the topic needs it: an ODE hands you the slope and asks for the path. Also written $\dot{y}$ (a dot on top = "derivative in time").

8. The dot notation $\dot{y}$ and $\ddot{x}$

Picture a swing: $x$ is how far it is displaced, $\dot{x}$ is how fast it's moving, $\ddot{x}$ is how hard it's being tugged back. That is the Harmonic oscillator the parent solves, $\ddot{x}=-{\omega }^{2}x$.

Why the topic needs it: second-order ODEs use $\ddot{x}$; you must rewrite them as two first-order equations because solvers only eat first-order ($\dot{}$, one dot) systems.

9. The step size $h$ and index $n$

Figure 4: starting dot $y_n$, follow the tangent slope for a width $h$, land at $y_{n+1}$. The gap between that landing point and the true curve is the step error — smaller $h$, smaller gap.

Why the topic needs it: the update rule $y_{n+1}=y_n+hf(t_n,y_n)$ (Euler) and the fancier Runge-Kutta methods are all written in these symbols.

10. The starting value $y_0$ and $y(t_0)=y_0$

Same speed-recipe but different starting spots give different journeys — like two cars obeying the same speed law but starting on different streets. The pair $(t_0,y_0)$ chooses one.

Why the topic needs it: this is the "Initial" in Initial Value Problem; solve_ivp(f, [t0,tf], y0) demands it.

11. $\omega$ and $\xi$ — the last two Greek guests

Why the topic needs it: $\omega$ appears in the oscillator $\ddot{x}=-{\omega }^{2}x$; $\xi$ appears in Simpson's error $-\frac{h^5}{90}{f}^{(4)}(\xi )$ — it just says "evaluated at some hidden point".

The prerequisite map

The left branch (function + interval + slivers + sum) feeds quad. The right branch (slope + step + start point) feeds solve_ivp/odeint. Both meet at the module. The connective tissue is the Fundamental Theorem of Calculus: it says accumulating slope is the same act as computing area — which is exactly why one module holds both.

Equipment checklist

Cover the right side; can you answer before revealing?