Energy-aware High Performance Multi-objective Optimization in Heterogeneous Computer Architectures. Applications on Biomedical (e-hpMOBE)
REPOSITORY OF THE PROJECT hpMoon (TIN2012-32039)
This repository (https://github.com/hpmoon) provides access to the most relevant codes of the project hpMooN (High Performance Multi-objective Optimization and Applications on Neuroengineering and Rehabilitation Technologies).
It contains three directories:
hpmoon
Including Matlab codes for multi-objective feature selection for both supervised and unsupervised classification problems related with BCI tasks.
The directories included in “hpmoon” includes a word file explaining the structure of the code and the way to use it.
A dataset file is also included to check the execution of the codes and to describe the format of the data file used by the code.
The toolboxes NSGA2 and somtoolbox are required by the codes. They are provided in the directory hpmoon_par, although they can be downloaded elsewhere (http://www.cis.hut.fi/projects/somtoolbox/ and http://www.iitk.ac.in/kangal/).
hpmoon-islands
Including distributed multi-objective algorithms for multi-objective optimization with high-dimensional decision spaces developed in the ECJ framework, a research Evolutionary Computation system written in Java (https://cs.gmu.edu/~eclab/projects/ecj/).
hpmoonOpenCL
Including OpenCL (https://www.khronos.org/) implementations for multicore and GPU platforms of a multi-objective feature selection in patterns with a high number of components (multi-objective optimization in high-dimensional decision spaces).