System architecting is a crucial early stage of the engineering design process: it establishes the high-level layout of a system. However, traditional architecting approaches often only consider a limited set of alternatives. This paper investigates strategies to integrate architectural decisions more tightly with downstream design optimization, enabling a thorough exploration of the hierarchical design space. Three optimization strategies are evaluated: global exploration, nested optimization, and decision chain optimization. These are tested on problems of varying complexity, from airfoil shape optimization to multi-objective benchmark problems to an industry aileron design study. The global exploration strategy optimizes all variables in a single loop and generally demonstrates the best performance. However, the nested strategy, which partitions variables into separate optimization loops, can be advantageous when a complete problem formulation is unavailable. Although the decision chain strategy using reinforcement learning provides interesting perspectives, it exhibits high complexity in implementation without significant performance gains. The main recommendation is to employ the global exploration strategy when possible, with nested optimization as an alternative if the complete design space is unknown. This work provides insights into optimization strategies connecting system architecting and concept development in engineering design.@en