3 ms·
My once-a-decade opportunity to post this thing that I wrote: https://peteshadbolt.co.uk/posts/ga/ https://peteshadbolt.co.uk/posts/ga/ Now back to life thank
by pjs_ 3y ago
My once-a-decade opportunity to post this thing that I wrote:
https://peteshadbolt.co.uk/posts/ga/ https://peteshadbolt.co.uk/posts/ga/
Now back to life thanks to Ruffle.
- juunpp 3y agoI think it's interesting and deserves a longer dive, particularly in car representation and crossover/mutation functions, as well as: > The optimality of a particular candidate solution (the fitness function) is determined by how far it travels before a) a mass touches the ground or b) time runs out. How do you sort among the millions of possible solutions that automatically die upon touching ground.
- HanClinto 3y agoThis was great, thanks! Really enjoyed this
- lemper 3y agobro, your cert is expired. just a friendly reminder.
- godelski 3y ago> The certificate for peteshadbolt.co.uk expired on 11/21/2023. What are the odds of that?
- riverdweller 3y agoApproximately 1.11%. Certificates issued by LetsEncrypt (the CA for this certificate) are valid for 90 days. Most users automate their renewal.
- godelski 3y agoLOL well thanks for doing the math!
- pjs_ 3y agoThanks! Fixed :)
- animal531 3y agoDo you have some resources for someone who wants to start to build their own version and/or some other GA?
- laurent_du 3y agoCan't stop watching.
- chriswarbo 3y agoI also made some pages (around a decade ago) which implement various optimisation algorithms, including GA (with just mutation, with just crossover, and with both). The task they're solving is less interesting though: find the brightest point on a randomly-generated image. In fact, being so low-dimensional, the enumeration algorithm is probably the best choice! http://www.chriswarbo.net/projects/optimisation http://www.chriswarbo.net/projects/optimisation