|
AMPL is a comprehensive and powerful
algebraic modeling language for linear and nonlinear optimization problems,
in discrete or continuous variables. AMPL lets you use common notation and
familiar concepts to formulate optimization models and examine solutions
AMPL models are analyzed in John Chinneck's
mProbe
software which collects statistics on their variables and constraints and
attempts to empirically verify the definition of convexity.
Michael Grant builds
disciplined convex programs from basic atoms.
Structural analysis of AMPL models has been analyzed in the past from
the point of view of
partial separability,
in order to convey such information directly to a solver.
Details pertaining to
the implementation of automatic differentiation in AMPL is described in
[Gay, 1991]
and
[Gay, 1996].
|