image-evol

An experiment in writing a genetic evolution algorithm that runs in wasm in the browser.

Curious? There is a presentation covering my learnings while working on this, and a deployed site that hosts the various demos that show different points of progress.

todos

• write the similarity code

• write mutation code

• write loop to mutate

• try to make mutate faster -- mutate in-place (this worked -- way, way faster)

• try to make mutate faster -- see how it changes with many more line segments -- no problem with even as many as 500 segments

• add filled triangles

• replace line segments with triangles

• colors & transparency

• implement population selection - mutation, tournament selection

• create a pool of mutation candidates and select the best ones at each iteration -> improve the convergence

later todos

• remove unneeded nalgebra crate?

• load an image as a reference rather than random

• UI controls to modify mutation rates

• npm run start -- Serve the project locally for development at http://localhost:8080.

• npm run build -- Bundle the project (in production mode).