From Theory to Practice with Particle Swarm Optimization, Using Python
There is a joke that cracks me up:
"Did you know that, before the clock was invented, people had to actively roam around and ask people the time?"
There is obviously no need to explain the joke, but if we were to overthink it a little bit (like good mathematicians do) we can say that the joke is about the fact that the information of a particle of a group can be used to inform all the other particles. This concept is actually way deeper than the joke I just said and can be exploited further.
Let's consider a self-organized system, such as bird flocking or fish schooling. We can define this system as one made of particles (e.g. a particle is a bird). We can also assume with a good degree of approximation that these particles move around in space adjusting their positions based on two factors:
- The best position that the specific particle knows: what the bird thinks is best for themselves.
- The global best position that is given by all the particles "communicating" with each other: what the bird is instructed to do by the "main bird"
Now, what is "best" in Nature? What is the best thing to do for the bird? What is the best for the "swarm"? I'm absolutely not the right person to ask this, as I really have no idea. What I DO know is that by observing this kind of behavior in Nature, we are able to formalize a very fascinating optimization algorithm. In other words, if we do define what is best, then we can use this evolutionary approach to optimize the function that we selected.
This algorithm is known as Particle Swarm Optimization (PSO). I know, this is a pretty big leap. What is "optimization"? Why are we talking about math all of a sudden? What is that we are optimizing? Throughout this article, I will try to cover all these steps and more importantly, we will use object-based programming in Python to create our own ParticleSwarmOptimizer() class. In other words, we will cover the PSO world from theory to practice.
Let's get started!
