lucasmpavelski / elmoead

ELM surrogate models for MOEA/D algorithm

Overview

This folder contain the source code for ELMOEA/D (Extreme learning surrogate models in multi-objective optimization based on decomposition). The src folder contain the ELMOEA/D source files and the lib folder its dependencies.

Building Instructions

The project is build with CMake version 2.8>. Other dependencies are a C++ compiler compatible with C++11 standard and gnuplot if plotting callback is used.

To build the executables execute the following in elmoead_benchmarks folder (assuming a Unix like system):

$ mkdir build && cd build $ cmake .. $ make

This will build the executables:

moead_surrogate - ELMOEA/D algorithm moead_alg - Non-surrogate MOEA/D algorithm test_all - unit tests

Running ELMOEA/D

To use ELMOEA/D run following command:

./moead_surrogate config_file.cnf final_pop.dat

where config_file.cnf is the parameters files and final_pop.dat is the final population objectives data.

The configuration file have the following syntax: =

The supported parameters are:

• seed <long>: random number generator seed
• callback_type: callback function called every iteration (NONE, PLOT, LOG)
• prob_name <string>: problem name (WFGn, ZDTn, UFn, DTLZn etc)
• prob_no_vars <long>: number of variables
• prob_no_objs <long>: number of objectives
• hyp_ref <real> <real> ...: hypervolume reference point (prob_no_objs real variables)
• de_operator <string>: differential evolution algorithm (DE/RAND/1/BIN, DE/RAND/2/BIN, DE/NONLINEAR)
• de_cr <real>: differential evolution algorithm crossover rate
• de_f <real>: differential evolution algorithm scaling parameter
• weight_type <string>: moea/d weight generation (UNIFORM, RANDOM)
• moead_type <string>: moea/d type (MOEA/D, MOEA/D-DE, MOEA/D-STM)
• moead_aggr_func <string>: moead aggregation function (TCHEBYCHEFF, INVERTED_TCHEBYCHEFF PENALTY_BOUNDARY_INTERSECTION etc)
• moead_no_partitions <long>: number of partitions in moead weight set
• moead_no_neighbors <long>: number of individuals per neighborhood
• moead_no_updates <long>: number of moead updates per generation
• moead_real_b <real>: moea/d-de chance to choose from neighborhood
• moead_sbx_xover <read>: moead sbx crossover rate
• moead_sbx_eta <read>: moead sbx eta parameter
• moead_pm_mut <read>: moead mutation rate
• moead_pm_eta <read>: moead polynomial mutaion eta parameter
• elm_act_func <string>: elm activation function (SIGMOID, RBF, GAUSSIAN, MULTIQUADRIC etc)
• elm_C <real>: elm regularization parameter", 1.0);
• elm_no_hidden <long>: elm number of hidden neurons
• elm_norm_input <boolean>: normalize input variables flag
• elm_norm_output <boolean>: normalize output variables flag
• no_evals: maximum number of real evaluations
• no_sel_partitions: elmoead number of partitions for selection weights
• archive_eps: elmoead archive minimum difference

If the parameter is not provided in the configuration file, a default value will be set and printed on the execution.

Disclamer

I do not own any of the libraries contained in the lib/ folder. Please respect their respective licences.

If you use this code please cite the paper:

• Lucas M. Pavelski, Myriam R. Delgado, Carolina P. Almeida, Richard A. Gonçalves, Sandra M. Venske, Extreme Learning Surrogate Models in Multi-objective Optimization based on Decomposition, Neurocomputing, Available online 6 November 2015, ISSN 0925-2312, http://dx.doi.org/10.1016/j.neucom.2015.09.111. (http://www.sciencedirect.com/science/article/pii/S0925231215016094)

This work was also based on the papers (please also cite them if relevant):

• Saúl Zapotecas Martínez and Carlos A. Coello Coello. 2013. MOEA/D assisted by rbf networks for expensive multi-objective optimization problems. In Proceedings of the 15th annual conference on Genetic and evolutionary computation (GECCO '13), Christian Blum (Ed.). ACM, New York, NY, USA, 1405-1412. DOI=http://dx.doi.org/10.1145/2463372.2465805.
• H. Li and Q. Zhang, Multiobjective Optimization Problems with Complicated Pareto Sets, MOEA/D and NSGA-II, IEEE Trans on Evolutionary Computation, vol. 12, no 2, pp 284-302, April/2009.
• G.-B. Huang, H. Zhou, X. Ding, and R. Zhang, “Extreme Learning Machine for Regression and Multiclass Classification,” IEEE Transactions on Systems, Man, and Cybernetics - Part B: Cybernetics, vol. 42, no. 2, pp. 513-529, 2012.