The thesis tackles the problem of structural design through topology optimization methods. The main contributions are the following. Firstly, a numerical benchmark is proposed considering the similarities of tests in the related literature, every test of the benchmark is well-established and used to compare the numerical and visual performance of the methods. Then, an analysis of the local optimization method Solid Isotropic Material with Penalization is carried out, in order to introduce two modications: 1) to reduce the computational cost in problems of stiness maximization, and 2) to minimize the volume of the structures. Finally, a novel technique to reduce the dimension of the problem through a proposal of a structure representation is introduced. This representation is well suited for mono and multi objective evolutionary algorithms in order to maximize stiness and minimize volume. All the algorithms take advantages of parallel computing to distribute the computational cost implied in the evaluation of solutions.