The method of modelling of the objective function in the problem of clusterization
 is described with the use of theory of combinatorial optimization. 
Some characteristic properties of this problem, the account of which allowed to generalize
 the concept of complication of its decision, are considered. It is shown that a presence 
in her of subclass of solvable problems depends on of a structure of input data, 
on the chosen estimation of similarity of objects and on the structure 
of argument of objective function, which laying out of element partitioning of sets into subsets. 
Introduction of subclass of solvable problems after a structure of input data 
conducted with the use of method of modelling their structure by the functions of natural argument, 
one the functions is combinatorial. It is shown that after the structure of argument 
of objective function of parallel of calculations in the cluster problem also taken to the solvable problems.