ECE 573 Advanced Engineering Design by Optimization
The course is primarily concerned with constrained numerical optimization
techniques as applied to various engineering design problems. Here the term
"numerical optimization" is referred to a process of finding optimal numerical
values for a set of design parameters in the sense that these parameters would
minimize an objective function which represents the performance of the design. The
term "constrained" is referred to the class of optimization problems in which
certain constraints are imposed on the parameters involved.
The techniques to
be studied in the course include simplex and interior-point linear programming,
convex quadratic programming (QP), semidefinite programming (SDP), sequential
quadratic programming (SQP) and interior-point algorithms for nonconvex constrained
optimization. Case studies of these methods as applied to digital filter design,
robust model predictive control, optimal force distribution for robotic dextrous
hands, optimal synthesis of attenna array patterns will be presented.