[CCoE Notice] Masters Thesis Defense

[Khator, Suresh]

Master's Thesis Defense

ILLUSTRATION OF NONLINEAR ROBUST OPTIMIZATION MODELS IN ENGINEERING DESIGN

Bhanu Kiran Susarla

Tuesday, November 8, 2011, 9.00 A.M.
Engineering Building I, Room 208 (Mechanical Engineering Small Conference Room)

Adviser: Dr. Jagannatha R. Rao

Abstract


The main focus of robust optimization has been put on linear models with uncertain parameters for many years. However, many real-life optimization problems are nonlinear. A robust-optimization formulation for nonlinear programming has been recently introduced by Zhang.Y (2007), which deals with first order robustness. But the effect of designation of different variables as state and control on the optimization results is unanswered. In this study, the above robust-optimization formulation is first applied to a structural nonlinear optimization problem, and a spring-mass-damper system to understand its effectiveness. Later, a third case study in which the designer has to differentiate the state and control variables is discussed to understand its implications on the optimization results. Finally, the famous Golinski's speed reducer problem is used to show that even a very small uncertainty in parameters could significantly alter the optimal solution, and explain the advantage of robust optimization over design using factor-of-safety.


Keywords: Robust nonlinear optimization, Factor of Safety, Data perturbations.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20111103/41ae479a/attachment.html