[Previous] [Next]

SLPloadprob

Purpose

Loads the nonlinear problem into the Xpress-SLP optimizer

Synopsis

procedure SLPloadprob(Name:linctr | gexp [Options:string])

Arguments

Name	name of the objective function
Options	set to "n" to force the use of the linear SLP optimizer even if the objective is quadratic.

Example

The following examples of the SLPloadprob procedure illustrate the different types of objective which can be defined:

SLPloadprob(OBJ)     ! where OBJ is a linctr
SLPloadprob(OBJcol)  ! This can be used if OBJcol is a column whose
                     ! activity contains the value of the objective
SLPloadprob(quadexp) ! This can be used if quadexp is a gexp
                     ! containing a quadratic objective
SLPloadprob(nlexp)   ! If nlexp is a general (non-quadratic,
                     ! nonlinear) function, additional structure
                     ! will be created to allow solution using
                     ! the linear optimizer

Further information

Unless otherwised specified on the call to the optimization function, the linear optimizer will be used to solve all problems except those with a quadratic objective. If Name contains only quadratic (and possibly linear) terms, then the quadratic optimizer will be used by default.

Any objective can be structured into a linear objective function by modelling it in an equality constraint, and transfering the resulting value to the (linear) objective. For example, if OBJECTIVE is the required nonlinear objective, then:
create a free mpvar OBJ
set OBJ = OBJECTIVE
and then do SLPloadprob(OBJ)

An example of this technique is shown in the Xpress-SLP User Guide. If a non-quadratic, nonlinear, objective is provided, then a transfer structure of this type will be created automatically. If the "n" option is included, then the transfer structure will also be provided if the objective is quadratic.

SLPloadprob loads the nonlinear problem into the Xpress optimizer. The linear equivalent (loadprob) will load only the linear part of the problem.