In traditional designmodeling, multi-objective formulations have often been used to capture the trade-o s inherent in concurrent design. However, such models result in one large aggregate model which does not retain the natural hierarchy of decisionmakers. To address this shortcoming, in this thesis we study Stackelberg Games Strategies to model the decision-making protocols of concurrency in design. The resulting parametric deformation may lead to singularities even in the presence of smooth problem data. We examine the regularity conditions necessary to obtain the rst-order necessary conditions at the leader's level. To numerically compute strategies in general nonlinear bi-level models, we implement an implicit nonlinear programming solver using sensitivity derivatives at the lower level. We develop two design case studies as illustrative examples: (1) the design of a nonuniform solid disc which maximizes energy storage, and maximizes a measure of ease of manufacture, and (2) the design of a single-point cutting tool which minimizes the cost of operation, and maximizes the behavior characteristics of the tool. v
[1]
Anthony V. Fiacco,et al.
Introduction to Sensitivity and Stability Analysis in Nonlinear Programming
,
2012
.
[2]
Jacqueline Morgan,et al.
Approximate solutions for two-level optimization problems
,
1988
.
[3]
K. Kiwiel.
Methods of Descent for Nondifferentiable Optimization
,
1985
.
[4]
C. A. Tiahrt,et al.
Bifurcation problems in nonlinear parametric programming
,
1987,
Math. Program..
[5]
J. Cruz,et al.
On the Stackelberg strategy in nonzero-sum games
,
1973
.
[6]
Fryderyk E. Gorczyca.
Application of metal cutting theory
,
1987
.
[7]
M. C. Shaw.
Metal Cutting Principles
,
1960
.
[8]
T. L. Vincent,et al.
Optimality in parametric systems
,
1981
.