Root Finding & Minimization 📉

"If $f(x) = 0$, what is $x$?"
From high school algebra to the Higgs Boson.

Lecture 2-5

1. The Problem 🌱

Slides 1-3

Finding where a function crosses zero (Root) or hits bottom (Minimum) is fundamental.

  • Root Finding: Solve $f(x)=0$.
  • Minimization: Solve $f'(x)=0$ (where slope is zero).
  • Methods are surprisingly similar!
🏔️

"It's just walking downhill."

2. Root Finding Methods 🕵️

Slides 4-15

Method 1: Bisection 🍰

Robust but slow. Divide the interval in half.

Method 2: Brent's Method 🚀

Combines Bisection with Parabolic Interpolation. Much faster.

Method 3: Newton-Raphson 🍎

Uses derivatives. $x_{new} = x - f(x)/f'(x)$. Quadratic convergence.

Method 4: SciPy `opt.newton` 🛠️

The professional way.

3. Minimization 📉

Slides 16-25

Golden Section Search 🏆

Like Bisection, but for minimums. Uses the Golden Ratio.

Brent's Method (Min) 🐆

Parabolic interpolation for the minimum.

Multivariate Minimization (SciPy) 🌐

Finding minimum in 3D space ($x,y,z$).

4. Curve Fitting (The Higgs) ⚛️

Slides 26-35

Fitting theoretical models to data is just Minimizing the Chi-Square error!

$$ \chi^2 = \sum \frac{(y_{data} - y_{model})^2}{\sigma^2} $$

5. Genetic Algorithms 🧬

Slides 40+

What if the function is ugly, discontinuous, or has many traps? Gradient methods fail. We use Evolution: Selection, Crossover, Mutation.

🦍 ➡️ 🧍

Brute Force Scan (Slow)

Genetic Solution (Smart)