Genetic Algorithms Fun

I have something to say for those of you who don't believe in the Darwin's Theory of Evolution - evolution just works! I mean it really works. One can get perfect and exact solution of a problem from an ugly set of numbers which doesn't make any sense. Here is the perfect and simple example of this:

(more info about this project is here)

See? It goes almost directly to a goal, all you need is to have one (here we can speculate if God has a goal).

To make evolution work we need an information which describes a creature - an information we can modify and pass to another generations. Mother Nature made it for us - things we call 'genes'. And as long as we copy everything from the nature, we just copied this perfect approach also to solve for example NP-hard problems. And we call it Genetic Algorithms (or GA).

'Genetic Algorithms'

We used GA in pgRouting before for solving Travelling Salesperson Problem (or TSP), and we used very nice and good working GAUL library for this. And it worked like a charm!

Now we are facing another challenge - to extend pgRouting by adding the Vehicle Routing Problem (or VRP) solver using GA. SO, this is kind of an announcement about what is cooking now at Georepublic's kitchen. I just wanted you to share my excitement about GA and thank Kristin Berg Bergvinsdottir for her theses dedicated to this problem and its solution.


Comments