Research Article

Multiobjective Genetic Programming Can Improve the Explanatory Capabilities of Mechanism-Based Models of Social Systems

Table 2

Structures of the discussed models (structures of the reference model, model 24, model 25, and model 29).

NameThe reference modelModel 24Model 25Model 29

RoleSelectionMechONOFFONOFF
RoleSocialisationMechONOFFOFFOFF
RoleLoad0.180  InvolvedxMarital + 0.039  InvolvedxParent + 0.128  InvolvedxEmployment + 60.952  (1 − MaritalStatus)  InvolvedxParentInvolvedxEmploymentInvolvedxEmploymentInvolvedxMarital
RoleIncongruence0.333  DiffExpectancyMarital + 0.333      DiffExpectancyParent + 0.333  DiffExpectancyEmploymentDiffExpectancyMaritalDiffExpectancyMaritalDiffExpectancyMarital
RoleStrain0.5  RoleLoad + 0.5  RoleIncongruenceRoleLoadRoleLoadRoleIncongruence
OutModOutMod = (1 − 0.25357  RoleLoad) + 11.18756  EmploymentStatusRoleLoadRoleLoadEmploymentStatus
InModInMod = (1 − 23.786  RoleLoad) + 2.010  (MaritalStatus + ParentStatus)RoleLoadRoleLoadRoleLoad
probFirstDrink(ProbOppIn + ProbOppOut)  disposition  (1 + 1.104  RoleStrain)DispositionDispositionDisposition  ProbOppIn
probNextDrink(ProbOppIn + ProbOppOut)  Disposition  (1 + 9.100  RoleStrain)DispositionDispositionDisposition

Elements not affecting agent drinking are highlighted in italic.