1

in my problem, I want to minimize the sum of the square of a decision variable * LARGE_CONSTANT. The reason for square is to incentivize the solver to spread the decision variable around equally; if I have to use the panic variables, I want to use them equally across the set of locations.

Some code (admittedly not enough to reproduce), follows:

declarations  
SITE: set of string  
c_min_panic: array(SITE) of mpvar  
c_max_panic: array(SITE) of mpvar  
end-declarations

! Objectives.
Min_Panicking:= sum(s in SITE) ((c_min_panic(s) * 10000)^2)  
Max_Panicking:= sum(s in SITE) ((c_max_panic(s) * 10000)^2)

However, this gives error: Mosel: E-101...Incompatible types for operator (mpvar' ^ integer' not defined).
Removing the ^ makes the problem work fine.

I am shocked why I can't square this? In fact, I see examples in the documentation that look like successful attempts to square the objective function. For example, Page 186 of the FICO Xpress User Guide seems to do it:

! Objective: minimise the total squared distance between all points  
  TotDist:= sum(i,j in RN | i<j) ((x(i)-x(j))^2+(y(i)-y(j))^2

What am I missing!! Tearing my hair out Thank you...

Geoff
  • 35
  • 3

1 Answers1

1

The code snippet does not show which solver module has been loaded: if it employs the Xpress Optimizer module, that is, starts with a line like

uses "mmxprs"

you also need to load the module 'mmnl' in order to extend the constraint handling capability of the Mosel language to include quadratic expressions, so

uses "mmxprs", "mmnl"

Another option would be to load the Xpress Nonlinear module:

uses "mmxnlp"
Susanne Heipcke
  • 156
  • 4