Multi-Level Modeling for Engineering Design Optimization

Thomas Ellman   John Keane  Mark Schwabacher  Ke-Thia Yao
Department of Computer Science, Hill Center for Mathematical Sciences
Rutgers University, Piscataway, NJ 08855
{ellman,keane,schwabac,kyao}@cs.rutgers.edu

HPCD-TR-44

June 1996

Physical systems can be modeled at many levels of approximation.  The
right model depends on the problem or subproblem to be solved. In many
cases, a combination of models will be more effective than a single
model alone.  Our research investigates this idea in the context of
engineering design optimization.  We present a family of strategies
for using multiple models to optimize engineering designs.  The
strategies are useful when multiple approximations of an objective
function can be implemented by compositional modeling techniques.  We
show how a compositional modeling library can be used to construct a
variety of locally calibratable approximation schemes that can be
incorporated into the optimization strategies. We analyze the
optimization strategies and approximation schemes to formulate and
prove sufficient conditions for correctness and convergence.  We also
report experimental tests of our methods in the domain of sailing
yacht design. Our results demonstrate a dramatic reduction in the CPU
time required for optimization, with no significant loss in design
quality.