Interactive Comparator of Multiobjective Optimization algorithms

David F. Dorado Sevilla, Universidad de Nariño, San Juan De Pasto - Colombia 2018
Most research problems are posed by taking into account more than one objective simultaneously and consequently there is no a single solution for it, but a set of possible solutions.  Multiobjective optimization, depending on the problem's characteristics, maximizes or minimizes the objective functions to find Pareto optimal solutions.  There is a wide range of multicriteria optimization methods that facilitate the task of solving this type of problems, but the results vary according to the methods used and Pareto optimality is not always reached, therefore, choosing a method to solve certain objective functions  and find the best possible solutions, is an open problem. The best way to choose an adequate method is to perform several tests that allow to understand how the methods work and which are the best results according to the users needs, but this is a difficult task for untrained people in the application of multiobjective optimization methods.

In this thesis, it is proposed the implementation of a User Friendly Interface that allows an interaction with the performance measures of the optimization methods and with the graphical reprecentacion of the solutions found, based on variations in the evaluation parameters.For this purpose, two methods commonly used in the literature are defined as NSGA-II (Non-dominated Sorting Genetic Algorithm) and MOPSO (Multi-Objective Particle Swarm Optimization) and test functions along with different Pareto fronts are used  to give diversity in the evaluation of each method. The aforementioned will allow the user to reach their own conclusions.                                                                   
                                            Figure.1. MOPSO Searching method                                                    Figure.2. NSGA-II Searching method                


Comparative study of NSGA-II and PSO as multiobjetive optimization methods in problems with a convex Pareto optimal front.                           International Conference on Information Systems and Computer Science Quito-Ecuador, INCISCOS_2016. David F. Dorado Sevilla, Maria J. Bravo Montenegro, Fredy A. Guasmayán Guasmayán, Diego H. Peluffo Ordoñez.

See full paper. Pag_181


See full thesis


Figure. Interactive comparator for NSGA-II and MOPSO

This interface alows user evaluate NSGA-II and MOPSO performance. In the link below, you can find all the files necesari to run the interface. Pleas download all files in the same folder, and open the main Matlab file named as metodo_comparativo.m and run the script. 
For more information please e-mail your questions.