Implementation of the globally-convergent method-of-moving-asymptotes (MMA)
algorithm for gradient-based local optimization, as described in:

	Krister Svanberg, "A class of globally convergent optimization
	methods based on conservative convex separable approximations,"
	SIAM J. Optim. 12 (2), p. 555-573 (2002).

However, in the case of NLopt, because we have no constraints other
than the bound constraints on the input variables, MMA reduces to an
extremely simple algorithm.  This doesn't mean it's a bad algorithm --
it may be especially suited to topology optimization problems where
the optimal design variables are almost always slammed against their
constraints -- just that this implementation doesn't really do justice
to the power of MMA for nonlinearly constrained problems.

It is under the same MIT license as the rest of my code in NLopt (see
../COPYRIGHT).

Steven G. Johnson
July 2008