CFSQP (C version of the Feasible Sequential Quadratic Programming algorithm)
is a gradient-based optimizer which allows multiple objectives and
constraints.
It performs an optimization within a multi-dimensional bounding box
and returns the point with the best (lowest) measure of merit.
Constraints may be used, in which case points where a constraint
is violated (> 0) are considered unsuitable.
The source code for CFSQP is not distributed with the interface.
Additional information can be found at the
FSQP home page.