Effective Discovery of Intrusion Protection Strategies

摘要

Effectiveness of discovery of strategy knowledge is studied for problems where the space of hypothesis of solutions is specified by game trees and target solutions are discovered by methods capable of systematic acquisition of expert knowledge about them. A version of Botvinnik's Intermediate Goals At First algorithm is developed for strategy formation based on common knowledge planning and dynamic testing of the plans in the corresponding game tree. Applied to the intrusion protection problem the algorithm for a range of types of knowledge in form of goals and rules demonstrates strong tendency to increasing the efficiency of strategy formation with an increase in the amount of knowledge available to the system.