Genetic Algorithms and Machine Learning for Programmers

Your Mission: Make Small Random Steps

Let’s define Monte Carlo simulations first, then see how Brownian motion will diffuse particles out of a paper bag. Geometric Brownian motion will build on the first model, giving a simulation of stock prices. Then you can sneak in potential price jumps, giving three models to simulate.

These models are stochastic. When you built a genetic algorithm in Chapter 3, ​Boom! Create a Genetic Algorithm​ you used a deterministic model—you determined the exact path of the cannonball for a given angle and velocity. In contrast, a stochastic model has a random element. You don’t know in advance exactly what will happen; however, you can work out properties, including an average or expected result, and how much variation ...