Elevator Pitch
Genetic algorithms are a way to “evolve” solutions to a problem, similar to biological evolution. This often reveals solutions that humans would never have thought of, like NASA’s ST5 spacecraft antenna! I will show you how these work, and how to use them to solve multiple kinds of problems.
Description
Yes, that’s right, geneTic, not geneRic. Genetic algorithms are shortcuts to “evolve” solutions to hard problems, by mimicking real-world biological evolution. They find “good enough” solutions much more quickly than precise calculation or brute-forcing your way through the entire solution space. Their randomness often reveals great solutions that humans probably would never have thought of, such as NASA’s twisty ST5 spacecraft antenna, developed by a genetic algorithm in 2006!
This talk will explain the general concept, and then walk through some examples, explaining terms and pointing out tradeoffs and nuances along the way.
You will come away equipped with a powerful technique for solving problems that might otherwise be computationally intractable, including in ways no human would probably ever think of directly!
Notes
This talk is simple enough that novices can grok it, but it can be useful for devs of all levels.
I have done this talk for several conferences so far, and am scheduled to do more soon. You can see the videos in my playlist at https://www.youtube.com/playlist?list=PLMrm16n64Buag1–3P3xf9KnrH330nrPc .
Agenda: - What are genetic algorithms? - Why are they useful? - Brief history - Examples of pre-existing usage - Overview of how they work - Walkthrough of creating ones to solve different kinds of problems (e.g., knapsack, traveling salesman, and/or others), pointing out terms, tradeoffs, nuances, etc. along the way - FAQs - Conclusion/recap - Q&A
The code is currently in Ruby, but I could easily do versions in Elixir, Python, JavaScript, or even C.