Papers by Author: Xiao Lin Zhang

Abstract: Analytical Target Cascading (ATC) is a method to partition the optimization of a complex system into a set of subsystem optimizations and a single system optimization according to the structure of the complex system, and coordinate subproblems toward an optimal system design. The constructed new optimization problem owns a hierarchical structure, which better matches the real organization structure of complex system design, so the ATC method provides a promising way to deal with the complex system. For each design problem at a given level, an optimization problem is to minimize the discrepancy between its responses and propagated targets. In ATC, for feasibility of subproblems, the target-response pairs are translated into the relaxation terms in which the weight coefficients is used to represent the relative importance of responses and linking variables matching their corresponding target, and achieve acceptable levels of inconsistency between subproblems when top level targets are unattainable in the hierarchical decomposition structure. Furthermore, weighting coefficients influence convergence efficiency and computational efficiency so that the suitable allocation of weight coefficients is a challenge. This paper adopts the Quadratic Exterior Penalty Method to deal with the weight coefficients that achieve solutions within user-specified acceptable inconsistency tolerances. Meanwhile, the method prototype will be tested on a numerical example and implemented using MATLAB and iSIGHT.

Abstract: ATC provides a systematic approach in solving decomposed large scale systems that has solvable subsystems. However， complex engineering system usually has a high computational cost , which result in limiting real-life applications of ATC based on high-fidelity simulation models. To address these problems, this paper aims to develop an efficient approximation model building techniques under the analytical target cascading (ATC) framework, to reduce computational cost associated with multidisciplinary design optimization problems based on high-fidelity simulations. An approximation model building techniques is proposed: approximations in the subsystem level are based on variable-fidelity modeling (interaction of low- and high-fidelity models). The variable-fidelity modeling consists of computationally efficient simplified models (low-fidelity) and expensive detailed (high-fidelity) models. The effectiveness of the method for modeling under the ATC framework using variable-fidelity models is studied. Overall results show the methods introduced in this paper provide an effective way of improving computational efficiency of the ATC method based on variable-fidelity simulation models.