Abstract :
Optimization problems involving multiple criteria are commonly found in a nuclear reactor design. For example, the focus on economical or safety aspects may lead to different reactor configurations. Solutions, which improve safety, may not lead to economical designs. Aiming to deal at same time with multiple criteria in reactor designs, we have developed a multiobjective genetic algorithm (MOGA) using concepts of Pareto optimality and niching techniques. Here, intended to show the advantages of using the MOGA, we applied it to a simplified two-criterion reactor core optimization problem. Using a simplification of a real-world problem, the computational cost associated to the reactor simulation could be reduced and exhaustive experiments could be done. In such experiments the MOGA could be compared not only with a standard genetic algorithm (SGA) but also with a brute force method in which the solutions search space was scanned. The obtained results have shown that the use of MOGA in such kind of problem enhances the quality of the optimization outcome, providing a better and more realistic support to the nuclear engineer decision.
