Decision variables

Decision variables are represented in the Xpress Kalis Mosel module by the types cpvar and cpfloatvar. They correspond respectively to finite domain variables taking their values in a given interval, or set of integers, called a domain and to continuous domain variables represented by real valued intervals. Conceptually the variables can be represented in the following way:

Figure 2.1: Conceptual representation of finite domain variables

Figure 2.2: Conceptual representation of continuous variables

Decision variables are declared by using the standard Mosel syntax. For instance, the following code extract shows how to create a finite domain variable my_cpvar, a static array of cpvar cpvar_array, and a dynamic array of cpfloatvar dyn_cpfloatvar_array:

declarations
   my_cpvar           : cpvar
   cpvar_array        : array(1..10) of cpvar
   dyn_cpfloatvar_array : array(range) of cpfloatvar
end-declarations

Note that entries of dynamic arrays of domain variables (arrays declared as dynamic at their creation or arrays for which some or all of the index sets are not known at the declaration like dyn_cpfloatvar_array in the example above) must be created explicitly using the Mosel procedure create:

declarations
  dyn_cpfloatvar_array : array(range) of cpfloatvar
end-declarations
forall(i in 3..30) create(dyn_cpvar_array(i))

After its declaration, the second step in the creation of a domain variable consists of defining its domain with the procedure setdomain.

Domain variables are also used in the definition of constraints and search strategy.

kalis defines a set of functions for accessing and modifying cpvar and cpfloatvar states: