genx provides modular building blocks to run simulations of optimization and search problems using Genetic Algorithms (GA).
The vision for genx is to be a flexible and greatly extensible library for implementing genetic algorithm applications. genx is written in Rust. The library's API utilizes functional programming paradigm and exposes it's API in that manner only.
The implementation is split into building blocks which are all represented by traits. This crate provides most common and probably all possible implementation for all building blocks. So it can be used for many problems out of the box.
Here's a trivial example that returns back individual selected using stochastic universal sampling
use genx::selection::stochastic_universal::stochastic_universal_selection; let num_parents:usize = 10; let fitness_values = vec![2.4,8.4,3.2,9.4,9.0,11.0,4.5,0.6,4.4,2.3,5.6,10.0,0.2,9.0,4.8,7.7]; let result = stochastic_universal_selection(&fitness_values, num_parents, None) .iter() .map(|&a| fitness_values[a]) .collect::<Vec<f32>>();
stochastic_universal_selection takes in the
fitness_value vector, number of parents it needs to select and a seed value which is
Option<u64>. It returns back the indices of selected individuals which we map to actual fitness values.
Mutation function takes in a single individual, distribution index and returns in the mutated individual using polynomial mutation for real valued individual.
use genx::mutation::polynomial::polynomial_mutation; let individual = 29.11; let result = polynomial_mutation(individual, 4.2, 4.0, None);
The returned value may or may not be equal as is mutated based on a randomly generated value, which for deterministic results can be seeded.
The genetic algorithm needs a population that it evolves with each iteration. A population contains a number of individuals. Each individual represents a possible candidate solution for an optimization problem for which the best solution is searched for.
A Genetic Algorithm proceeds through following operations:
- Encoding: Binary, Real Values, Order, Tree, etc.
- Selection: Selecting individuals after fitness evaluation.
- Crossover: Creating new individuals from selected pool of individuals.
- Mutation: Making stark changes in generated individual for diversification.
- Convergence: Test for goal accomplishment or convergence.
The building blocks currently available (defined as traits) are:
This crate provides multiple implementations for each one of those operators. So one can experiment with combining the different implementations to compose the best algorithm for a specific search or optimization problem.
Add this to your
[dependencies] genx = "0.3.0"
If you are not using Rust 2018 edition add this to your crate root:
extern crate genx;
Why Genetic Algorithms
Genetic Algorithms are at the core of soft computing which is a branch of computing that comes to rescue when problem at hand is not feasible to be solved using hard computing. There are several advantages of genetic algorithms:
- Algorithms are adaptive and can adjust to the change of dynamic environment
- The approach makes use of approximate solutions to solve problems that may be either unsolvable or too time-consuming to solve with current hardware.
- Imprecise but usable solutions to complex computational problems allowing researchers to approach some problems that traditional computing cannot process.
I started this project mainly to learn about genetic algorithms (GAs) as part of my college curriculum where am studying severals methods for soft computing. I found myself looking for simple easy to use modular functions that I can use to learn more about each step of GA.