felixboelter / large-scale-waste-management-optimisation

Extension of the waste supply chain using multi-objective evolutionary algorithms.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Large-scale multi-objective optimisation for sustainable waste management using Evolutionary Algorithms

A meta-heuristic approach to sustainable waste management.

Installing the prerequisites

Make sure to append the ibmdecisionoptimization and conda-forge packages to the current channels. Write this into a terminal:

$ conda config --append channels ibmdecisionoptimization`
$ conda config --append channels conda-forge

Installation

Clone the repository and install the requirements file via conda.

$ git clone https://github.com/felixboelter/large-scale-waste-management-optimisation
$ cd large-scale-waste-management-optimisation
$ conda create --name <env> --file requirements.txt

Usage

The main.py file includes a class which generates and solves all of the graphs using the algorithms found in the source folder.

You can change the crossover and mutation probabilities by changing the crossover_probs and mutation_probs parameters with a list of selected probabilities between 0 and 1.

Results

Solutions to every solved graph should be found in a created Results folder. Which will include:

  • A csv file with the numeric solution of the best hypervolume solutions for each algorithm.
  • The solved graph of the best hypervolume solution for each algorithm.
  • A numpy file (.npy) containing all the solutions of the algorithms and decision values.

About

Extension of the waste supply chain using multi-objective evolutionary algorithms.

License:GNU General Public License v3.0


Languages

Language:TeX 50.5%Language:Python 49.5%