Using a Genetic Algorithm to Optimize Developer Conference Schedules

Everything about organizing an event is hard. Finding speakers, estimating attendance, securing a venue, scheduling, letting people know about the event, getting them badged, getting them wifi, getting them fed… all of this is hard.

One of the most outrageous-sounding solutions I’ve ever heard when it comes to organising an event is the brain child of Filip Hracek. Of all things, he proposed using a genetic algorithm to organise his conference schedule. 

Taking into account all of the different stages of his developer conference, he calculated that there were a whopping 23! (factorial, that is) permutations of how the schedule could be lain out, taking into account talks, breaks and lunches across a single track and two days. He then went on to say, assuming considering each variation took a microsecond, that an organiser could potentially take 819 million years to go through all the variations.


People usually use their brains to come up with a schedule that works based on the inexact science of common sense. However, by the power of automation, genetic programming, and (one can only assume) brimstone, he seems to have landed on the perfect timeline for a developer conference.

For the full article, go here: