Skip to content

litchipi/Petridish

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

89 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Petridish

A library to create a Genetic Algorithm library, written in Rust. Examples are located in the example directory. Meant to be used inside a Rust project after defining own structs implementing mandatory traits, has some bindings to Python using pyo3 (see examples & experiment with them)

Principle

A Genalgo struct gathers datasets and feed them into a Lab.

A Lab is a set of Algo, which manages and feed data to Cells.

After all data of all datasets fed the cells, the cells are passed into a GenalgoMethod to prepare the next generation, and get the best cells.

The GenalgoMethod uses MutationProcess and BreedingMethod to make operations on Cell's genomes

Each algo's best cell is injected into the population of another algo, based on a "lab map".

The algo defined as "output algo" is the one used to get the final optimised genome.

Philosophy

Everything must be moddable using traits. Are traits:

  • Cell
  • Algo
  • GenalgoMethod
  • MutationProcess (Work in progress)
  • BreedingMethod (Work in progress)
  • LabMapFormat (Used to assist LabMap creation)

Implemented Optimisation methods

  • Darwin (Mix of technics, including CMA on elites and natural selection. Custom made for experiments)
  • RandOpti (Get random genomes each generation, used for noise in LabMap)

Implemented BboB functions

On the code of example/benchmarking/, you can test the optimisation efficiency using Black-Box Optimisation Benchmark functions (BBOB). The following functions are implemented:

A special thanks to mazhar-ansari-ardeh for his very high quality website about these functions

About

Genetic algorithm library written in Rust

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published