Many Objective Evolutionary Algorithm

A set of evolutionary algorithms to solve many-objective optimization. Hybridization between the algorithms are also facilitated. Available algorithms are: 'SMS-EMOA' 'NSGA-III' 'MO-CMA-ES' The following many-objective benchmark problems are also provided: 'DTLZ1'-'DTLZ4' from Deb, et al. (2001) and 'WFG4'-'WFG9' from Huband, et al. (2005) .

MaOEA

The goal of MaOEA is to facilitate easy hybridization of algorithms for many objective optimization. In the package, several algorithms are available: SMS-EMOA, NSGA-III, and MO-CMA-ES. Each of these algorithms can be accessed independently. Using the main function, the algorithms can be called for specific number of iterations. Alternatively, if the hybridization follows a more complex rule, users may prefer to call the algorithm directly in their optimization loop. This will call the algorithm (i.e., the offspring generation and selection scheme) for a single iteration.

The package uses PyGMO (https://esa.github.io/pagmo2/) to compute hypervolume and hypervolume contribution.

Installation

You can install the released version of MaOEA from CRAN with:

install.packages("MaOEA")

Please note that MaOEA requires the users to have installed Python (see https://www.python.org) and being able to use the PyGMO module. Installation instruction for PyGMO is available in https://esa.github.io/pagmo2/install.html. Users can also try to use the function provided in the package:

MaOEA::install_python_dependencies()