Hierarchical multilevel optimization for reliability target allocation in probabilistic design of decomposed systems

Analytical target cascading (ATC) is a methodology for translating system-level design targets to subsystem and component design specifications in hierarchically decomposed optimal system design problems. In previous work we extended the ATC formulation to account for uncertainties, where the bounds of the probabilistic design constraints were chosen arbitrarily and held fixed during the ATC process. In this work, we extend the probabilistic ATC formulation to include reliability targets in the vector of cascaded quantities. In this manner, we quantify the optimality-reliability tradeoffs for each element of the decomposed system and compute the probabilistic constraint bounds required to satisfy the overall system reliability target.