Initializing help system before first use

Nonlinear objectives

Xpress NonLinear works with nonlinear constraints. If a nonlinear objective is required (except for the special case of a quadratic objective — see below) then the objective should be provided using a constraint in the problem. For example, to optimize f(x) where f is a nonlinear function and x is a set of one or more variables, create the constraint

f(x) - X = 0

where X is a new variable, and then optimize X.

In general, X should be made a free variable, so that the problem does not converge prematurely on the basis of an unchanging objective function. It is generally important that the objective is not artificially constrained (for example, by bounding X) because this can distort the solution process. Also, as such an objective transfer row is not a real constraint, no error vectors should be added (row can be enforced); feasibility should be provided by the transfer variable X being free.