Enhanced Collaborative Optimization: A Decomposition-Based Method for Multidisciplinary Design

Abstract

Astute choices made early in the design process provide the best opportunity for reducing the life cycle cost of a new product. Optimal decisions require reasonably detailed disciplinary analyses, which pose coordination challenges. These types of complex multidisciplinary problems are best addressed through the use of decomposition-based methods, several of which have recently been developed. Two of these methods are collaborative optimization (CO) and analytical target cascading (ATC). CO was conceived in 1994 in response to multidisciplinary design needs in the aerospace industry. Recent progress has led to an updated version, enhanced collaborative optimization (ECO), that is introduced in this paper. ECO addresses many of the computational challenges inherent in CO, yielding significant computational savings and more robust solutions. ATC was formalized in 2000 to address needs in the automotive industry. While ATC was originally developed for object-based decomposition, it is also applicable to multidisciplinary design problems. In this paper, both methods are applied to a set of test cases. The goal is to introduce the ECO methodology by comparing and contrasting it with ATC, a method familiar within the mechanical engineering design community. Comparison of ECO and ATC is not intended to establish the computational superiority of either method. Rather, these two methods are compared as a means of highlighting several promising approaches to the coordination of distributed design problems.

Multidisciplinary and Multilevel Aircraft Design Methodology Using Enhanced Collaborative Optimization

“…Another major difference from the works presented to date, 11,12,18 is that the compatibility parameter is a dynamic quantity. This is because it has been found that an early compatibility between subspaces may jeopardize the actual multilevel optimization problem.…”

“…6,12,[14][15][16][17] In ECO, the system level optimization is simply an unconstrained minimization problem. The global objective function (i.e.…”

On the Application of Differential Geometry to MDO

“…10 Some solutions have been proposed to keep the concept underlying the CO architecture while overcoming these issues: Modified Collaborative Optimization (MCO) from DeMiguel and Murray 25 and Enhanced Collaborative Optimization (ECO) from Roth and Kroo. 26 DeMiguel and Murray provide equivalence and convergence proofs for their proposal, 27, 28 but it is interesting to note how similar their formulation ends up being to that of ATC. This is meant not as a disparagement of their work but rather as an observation that certain solution methods (i.e.…”

Multi-modality in augmented Lagrangian coordination for distributed optimal design

“…In fact, the difficulties associated with disconnected feasible domains have been observed for ATC by Kim and Papalambros (2006). Other coordination approaches that use alternating minimization such as MDOIS (Shin and Park 2005), BLISS (Sobieszczanski- Sobieski et al 2000), and enhanced collaborative optimization (Roth and Kroo 2008) may also be affected by the behavior of the sequential process.…”