<p>This paper proposes an engineering method for optimizing component type selection, topology and size of complex structures composed of rods, beams, plates, or their combinations. A generalized problem formulation for structural component type selection, topology and size optimization is proposed. Two key types of constraints, component type uniqueness constraint and component association constraint, are defined, which distinguish the component type selection optimization problem from topology optimization problem of complex structures. Based on a two-level optimization strategy, a modified genetic algorithm and the dual method are employed to comprehensively optimize component types, topology and size of complex structures by extending the approximation concept. The mathematical model and solution strategy for this problem are detailed, and the effectiveness and efficiency of the proposed method are verified through numerical examples. The results demonstrate that this method can simultaneously optimize component types, structural topology and size, achieving a lighter structural design compared to traditional topology optimization. Moreover, the number of required structural analyses is significantly reduced. This study provides a comprehensive and efficient solution for the design of complex engineering structures, offering a valuable approach for optimizing component types, topology and size in engineering applications.</p>