This work verses about contributions in the optimization of laminated fiber reinforced composites. At first, it is presented a new solution concept to Discrete Material Optimization (DMO) problems based on Sequential Approximate Optimization (SAO) techniques, employing new compliance approximations in terms of intermediate variables which are explicit functions of the DMO weights. The proposed methodology improves significantly the quality of the approximations and consequently the general convergence characteristics of several DMO compliance minimization (stiffness maximization) problems. Then, it is presented a Hierarchical Optimization scheme for laminated composite structures, considering simultaneous design of macroscopic (structural) and microscopic (material) levels. The macroscopic level takes into account orientations and fiber volume fractions of unidirectional composite layers. The microscopic level considers the cross-sectional size and shape of the reinforcement fibers, assuming them elliptical. Both levels are coupled and the objective is to minimize compliance under a total fiber volume fraction constraint. It is shown that changes in the shape of the fibers permit to increase structural stiffness. An assessment of microstructural stresses is also included and shows the fibers'; shape influence on layers'; stress concentrations.