ev-MOGA Multiobjective Evolutionary Algorithm

Grupo de Control Predictivo y Optimizacin Heurstica http://ctl-predictivo.upv.es Universitat Politcnica de Valncia (SPAIN) Contact: juaherdu@isa.upv.es (Juan M. Herrero) ev-MOGA Multiobjective Evolutionary Algorithm has been developed by the Predictive Control and Heuristic optimization Group at Universitat Politcnica de Valncia. ev-MOGA is an elitist multi-objective evolutionary algorithm based on the concept of epsilon dominance. ev-MOGA, tries to obtain a good approximation to the Pareto Front in a smart distributed manner with limited memory resources. It also adjusts the limits of the Pareto front dynamically. Details about ev-MOGA are described in (please, cite this algorithm as): [1] M. Martnez, J.M. Herrero, J. Sanchis, X. Blasco and S. Garca-Nieto. Applied Pareto multi-objective optimization by stochastic solvers. Engineering Applications of Artificial Intelligence. Vol. 22 pp. 455 - 465, 2009 (ISSN:09521976). The algorithm is also described in: [2] J.M. Herrero, M. Martnez, J. Sanchis and X. Blasco. Well-Distributed Pareto Front by Using the epsilon-MOGA Evolutionary Algorithm. Lecture Notes in Computer Science, 4507, pp. 292-299, 2007. Springer-Verlag. (ISSN: 0302-9743) ev-MOGA has been used in: [3] J.M. Herrero, X. Blasco, M. Martnez, C. Ramos and J. Sanchis. Robust Identification of a Greenhouse Model using Multi-objective Evolutionary Algorithms. Biosystems Engineering. Vol. 98, Num. 3, pp. 335 - 346, Nov 2007. (ISSN 1537-5110) [4] J.M. Herrero, X. Blasco , M. Martnez, J. Sanchis. Multiobjective Tuning of Robust PID Controllers Using Evolutionary Algorithms. Lecture Notes in Computer Science, 4974, pp. 515 - 524, 2008. Springer-Verlag. (ISSN: 0302-9743) [5] J. M. Herrero, S. Garca-Nieto, X. Blasco, V. Romero-Garca, J. V. Snchez-Prez and L. M. Garcia-Raffi. Optimization of sonic crystal attenuation properties by evMOGA multiobjective evolutionary algorithm. Structural and Multidisciplinary Optimization. Vol. 39, num. 2, pp. 203 - 215, 2009 (ISSN:1615-1488). [6] G. Reynoso, X. Blasco, J. Sanchis. Diseo Multiobjetivo de controladores PID para el Benchmark de Control 2008-2009. Revista Iberoamericana de Automtica e Informtica Industrial. Vol. 6, Num. 4, pp. 93 - 103 , 2009. (ISSN: 1697-7912)

Basic instructions
The ev-MOGAdescription.pdf file contains the description of the ev-MOGA algorithm. You should read it before using the algorithm in order to understand how it works. Two multiobjective problems mop1.m and mop4.m are included as examples. Basic instructions: 1) Create the Matlab function used to evaluate the objective functions (e.g. mop1.m) 2) Modify run_evMOGA.m. which contains the parameter configuration of the evMOGA and defines the optimization problem to solve. 3) Execute the script run_evMOGA .m to run the ev-MOGA. After execution, variables ParetoFront and ParetoSet variables are obtained in the workspace.