The paper considers the models and algorithms of intelligent components intended for active analyzing. The offered approach is based on simulation of computer attacks by using attack graphs, logical trees and solving the probabilistic questions. This article discusses developing of the security expert system by using new methods for efficient and automated penetration testing. Methodologies of automated planning and partially observable Markov decision processes are suggested in order to improve results and build flexible security evaluation expert system.
Ключові слова: penetration testing, ethical hacking, artificial intelligence, expert system, planner, automation, attack graphs, pentesting system, algorithm, modelling, POMDP