An efficient design formulation for feedforward Active Structural Acoustic Control (ASAC) systems for complex structures and disturbances is presented. The approach consists in a multi-level optimization procedure. The upper level part is carried out in the modal domain, where the optimum modal control forces and modal error sensor components which minimize the total radiated power are obtained. These optimum modal parameters are then used in a set of lower level optimization problems to find the physical characteristics of the actuators and sensors to be implemented. In this work, the developed formulations are demonstrated in three systems of increasing complexity. First, a simply supported plate excited by at a single frequency is presented. The study of this relatively simple case serves as a benchmark for more complex systems permitting the evaluation of the performance of the design approach. Then, the formulation is implemented for the case of a simply supported cylinder under multiple frequency excitations. In this case numerical techniques are used to obtain the structural and acoustic responses showing the capabilities of the formulations to be implemented to complex structures. In both cases (i.e., the simply supported plate and cylinder), the results show that the optimum configurations yield significant reductions in the total radiated acoustic power depending on the number of control channels and allowed control effort. Moreover, it is demonstrated that the proposed design approach gives a clear insight into the relative contributions of the modes to the sound field providing a better understanding of how they have to be controlled in order to minimize sound radiation. Finally, the design approach is validated in an experimental arrangement for the attenuation of sound radiation from and enclosed box structure under a realistic periodic disturbance. Up to 36dB of attenuation in the acoustic power is obtained in the experiment demonstrating the effectiveness of the proposed design approach for practical applications.