MULTIAGENT SYSTEM MODEL FOR DETECTION OF POLYMORPHIC VIRUSES

Abstract

The paper establishes that multi-agent systems (MAS) are a powerful tool for detecting polymorphic viruses. Such a system uses several intelligent agents (IAs), each of which has its own specific role in the process of detecting and trawling polymorphic viruses. he concept of multi-agent systems (MAS) has become an important topic of interest in the field of artificial intelligence. The main advantages of using MAS for detecting polymorphic viruses are identified: parallel processing and efficiency; distributed threat detection; intelligent interaction between agents; adaptability to new threats; scalability; prediction and detection of anomalies; dynamic response to threats; distributed learning based on experience; reducing the load on one point of the system. A model of a multi-agent system (MAS) for detecting polymorphic viruses is proposed, which includes: a set of agents; a set of computer system states; a set of possible agent actions; a transition function between states; a reward function that evaluates the effectiveness of selected actions; a monitoring function that determines what information each agent receives; the probability of transitioning to a new state after the agents perform actions; an agent strategy that determines what action it chooses in each state. The intelligent agent of this MAS consists of the following modules: an analysis module, a module for classifying polymorphic viruses by complexity levels, and a decision-making module. Agents operate in an environment and can cooperate or compete. They operate according to an algorithm. The algorithm of the proposed MAS: information collection, detection of polymorphic viruses, classification of polymorphic viruses, decision making.