pymetagen-datalabupo 0.1.1

A Python Package problem optimization.

Metagen

Introduction

Metagen is a simple python library for solving optimization problems by metaheuristics algorithms. The project provides a set of already defined metaheristics and provide the basic tools to allow the user redefine/build its own algorithms in a simple way.

This project is currently maintained by David Gutierrez Áviles and Manuel Jesús Jiménez Navarro.

https://datalab.upo.es/

Link to the original paper: Metagen

Instalation

Metagen only requires python (>=3.10) as the library has been built completely on python.

For the instalation, the easiest way is to use pip:

pip install pymetagen-datalabupo

Tutorial

In orfer to work metagen we first need to import the Domain definition and an available metaheuristic.

from metagen.framework import Domain, Solution
from metagen.metaheuristics import RandomSearch

Every problem in metagen is defined by two elements: a domain and a fitness.

The domain of the problem is constructed by using the Domain imported previously . In this case we defined a simple domain with one integer variable in range [-10, 10].

domain: Domain = Domain()
domain.define_integer("x", -10, 10)

The fitness is a callable which receives a Solution as parameter and returns a float value.

def p1_fitness(solution: Solution) -> float:
    x = solution["x"] # The variable represents a potential solution instance which can be used as an map returning the builtin value.
    return x + 5

Finally, we only need to instantiate our metaherustic (in this case a simple random search) indicating the domain and fitness and run it returning the best solution.

random_search: RandomSearch = RandomSearch(domain, fitness)
solution: Solution = random_search.run()

Development

New contributors from all experience levels are welcomed. To contribute, you can open an issue or sending a pull request.

For testing the repository you just need to execute the following command:

pytest test

API reference

The official documentation is available in: https://pymetagen.readthedocs.io.

References

• cvoa: https://www.liebertpub.com/doi/10.1089/big.2020.0051