Collaborative optimization

From DDL Wiki

Collaborative optimization is a type of decomposition technique that was developed for large-scale design optimization problems that require multiple analysis routines to evaluate a solution. For example, it has been used to design a single-stage-to-orbit launch vehicle (Braun et al., 1996).

Collaborative optimization is designed to allow each discipline to solve its subproblem in parallel with the others. Auxiliary design variables are added to each subproblem. Optimizing the system-level problem determines design variable values that are then sent to the subproblem as target values. Each subproblem determines values for its local design variables in order to meet the targets as closely as possible subject to its local constraints. Sensitivity analysis of the subproblems provides gradients for the system-level problem, which improves the efficiency of the approach. The procedure iterates until it converges.

Decision-based collaborative optimization extends this idea to include design for market systems issues such as demand and profit. In Renaud and Gu (2006), the system-level objective is to maximize the expected utility of the net revenue.

References

Braun R.D., Kroo, I.M., and Moore, A.A., 1996, Use of the collaborative optimization architecture for launch vehicle design, AIAA-96-4018, Proceedings of the Sixth AIAA/NASA/USAF/ISSMO Symposium on Multidisciplinary Analysis and Optimization, Bellevue, Washington.

Renaud, J.E., and Gu, X., 2006, Decision-Based Collaborative Optimization of Multidisciplinary Systems, Decision Making in Engineering Design, K.E. Lewis, W. Chen, and L.C. Schmidt, eds., ASME Press, New York, pp. 173-186.