This page contains links to technical reports, software packages, presentation slides, and other useful materials. Feel free to browse on and to contact the authors / maintainers of each resource.
R package for sample size calculations in the experimental comparison of algorithms
This package is being actively developed by Prof. Felipe Campelo and Fernanda Takahashi. It contains routines for estimating both the number of instances and the number of runs/instance when comparing the mean performance of two algorithms for any given problem class.
R package for component-based design and analysis of multiobjective evolutionary algorithms based on decomposition
This package is being actively developed by Prof. Felipe Campelo and Prof. Claus Aranha (Tsukuba University, Japan), with theoretical support by Prof. Lucas Batista. The package contains a modular implementation of the MOEA/D framework, allowing for easier replication of existing variants and investigation of new ones.
Differential Evolution for Multiobjective Optimization
Github repository: https://github.com/ORCSLab/DEMO
Octave/Matlab implementation of the Differential Evolution for Multiobjective Optimisation algorithm and several variants:
- A posteriori methods (without preferences):
- A priori or interactive methods (with preferences):
- R-DEMO: R-NSGA-II (Deb et al., 2006) but using the DEMO instead;
- PBEA-DEMO: IBEA, but using a reference point (Thiele et al., 2009);
- PAR-DEMO(nds): Preference-based Adaptive Region-of-interest DEMO using non dominated sorting (Goulart and Campelo, 2015);
- PAR-DEMO(ε): Preference-based Adaptive Region-of-interest DEMO using indicators (Goulart and Campelo, 2015);
Loudspeaker design model
Toy problem for electromagnetic optimisation
Github repository: https://github.com/ORCSLab/Loudspeaker-model
This model is inspired on the loudspeaker design example originally proposed by Infolytica Corporation. The objective of this problem is the minimization of the total volume of material used, subject to the generation of a given magnetic flux density in the air gap defined by variable x9.
This problem can also be very easily turned into a multiobjective optimization example, by considering the minimization of volume and the maximization of the magnetic flux density.
This model was developed (a long time ago) by Prof. Felipe Campelo.
Design and Analysis of Experiments
Github repository: https://github.com/fcampelo/Design-and-Analysis-of-Experiments
This repository contains lecture slides, R codes, examples and other materials related to Prof. Felipe Campelo's course on Design and Analysis of Experiments, which is offered twice a year for the Graduate Program in Electrical Engineering.
R package for experimenting with Differential Evolution
This package is being actively developed by Moisés Botelho and Prof. Felipe Campelo, as part of the former's M.Sc. work. The package implements the basic DE algorithm introduced by Storn and Price in 1996 in a modular fashion, allowing for easily modifying the main components of the algorithm in order to perform comparative experiments and performance evaluation. The package currently provides around 20 different recombination operators that can be used with the DE, and will soon provide a variety of differential vector composition methods as well.
Introdução: Otimização e Programação Linear
(In Portuguese) Tutorial
Repositório Github: https://github.com/fcampelo/Otimizacao-Intro-Prog-Linear
Este repositório contém alguns materiais introdutórios relativos a conceitos de otimização e programação linear, originalmente desenvolvidos para a disciplina de Otimização Aplicada à Engenharia de Processos do Curso de Especialização em Automação Industrial (CEAI/UFMG).