Exploring multidisciplinary design optimization (MDO) architecture based on global optimization algorithm

Ref. 1 proposed MDF (Multi-disciplinary Feasible) architecture. Several researchers in Refs. 2 through 5 tried, in our opinion, to replace MDF with something else better unsuccessfully. We now return to MDF and explore how to improve it. Section 1 of the full paper compares the MDF architecture with the decomposition architectures used in Refs. 2 through 5, arriving at the conclusion that MDF has better computational performance. To enhance its computational performance, section 2 presents the requirements for problem solving in multi-disciplinary analysis (MDA). With these requirements, it converts the MDA solution into an optimization problem as shown in eq. (6). To find the solution to the problem, Section 2 combines the genetic algorithm(GA) with the complex algorithm(CA) to form a new GACA for solving the optimization problem, which is shown in Fig. 2. Using the GACA, section 3 improves the MDF architecture as shown in Fig. 3 and thus establishes a bi-level MDF architecture. The calculation results of a typical numerical example, given in Table 1, show preliminarily that the bi-level MDF architecture has much better computational performance than other architectures.