| Resumen |
Most of the optimization problems in engineering have multiple objectives that must be optimized at the same time. They are named multi-objective optimization problems. The solution for these algorithms is composed by a set of solutions (instead of only one single solution) called the Pareto front. The algorithms to solve them grow exponentially their complexity when the number of objectives increases. Actually, the most popular multiobjective Evolutionary Algorithms (such as NSGA-II, PISA, etc.) have difficulties approximating the optimal Pareto front when the number of objectives is greater than 3 (called many-objective problems). The hypervolume-based algorithms have demonstrated to be capable to find good solutions for many objective problems, however they requiere large execution time. A very recent algorithm Partition and Selection Algorithm (PSA) [3] proposes a simple mechanism to improve diversity mechanism in many-objective evolutionary algorithms. On the other hand, the reduction of objectives is a viable alternative to deal this problem. If the number of objectives is reduced without altering the solution (or until a certain acceptable error) the problem could be solved by the current well established algorithms, reducing the execution time and improving the approximation to the Pareto front. Additionally, the visualization/analisys of the final Pareto front is easier handled by the final Decision Maker |
