The aim of this chapter is to show the interest of considering scheduling problems for the design improvement of Flexible Manufacturing Systems (FMS). The problem occurs when we wish to reconsider the design of an FMS by evaluating its functioning using a low-level model. The experimental results we have obtained show that, in many cases, a simple reorganization of the production cells can improve the overall productivity of an FMS. This problem is known in the literature as the machine assignment problem. The low-level evaluation model we have considered permits to synchronize the handling material system with the production tools, in order to minimize the time required for the production of a given set of jobs (makespan). We propose a hybridized solution approach that combines different optimization strategies: integer linear programming, metaheuristics and discrete event approaches. Integer linear programming is used for elaborating an initial machine assignment (an optimal one for the high-level model). New machine assignments are constructed by ants, and evaluating thanks to a low-level model evaluation. The latter is a black box optimization subroutine that combines iterated local search and discrete event approach.