Verification of the security systems antagonistic agents behavior model

Abstract

Model verification is a very important step in the methodology for modeling the security systems antagonistic agents behavior in general and system dynamics in particular. By verifying the behavior model of antagonistic agents we mean a process that includes both formal/quantitative tools and informal/qualitative ones. The article presents the process of creating a model of antagonistic agents behavior. The assumptions underlying the model and the limitations of the created model are preliminarily formed. The components of the model are distinguished: a defender submodel, an attacker submodel and a confrontation environment submodel. For each of the submodels, the processes and relationships in models are described, the variables used for modeling are defined. Processes and relations between variables are presented in the form of a system of linear and differential equations. Based on the given system of equations of the mathematical model, a system-dynamic model of the interaction of antagonistic agents is constructed. It is shown that for the practical use of the software implementation of the behavior model, the verification procedure is mandatory. The main groups of tests that need to be performed using the model are listed to confirm its adequacy to the conditions of use and the goals for which it was developed. The results of testing the system-dynamic behavior model for the main group of verification tests at each of the three main stages of model verification are presented: structural tests, structure-oriented behavior tests and behavior model tests. Based on the results obtained, the special importance of structurally oriented behavioral tests is emphasized. These are powerful behavioral tests that can provide information about potential structural weaknesses. These tests seem to be the most promising area for research on model verification.