Abstract
The object of this research is artificial immune systems. The problem addressed in the study is improving the responsiveness of cyberattack detection in information systems while ensuring a predetermined level of convergence, regardless of the number of destabilizing factors. The subject of the research is the cyberattack detection process. A cyberattack detection method for information systems based on artificial intelligence technologies is proposed. The originality of the method lies in the use of additional enhanced procedures that allow: – initializing the initial population of swarm agents and verifying information system parameters using an improved bat algorithm, which minimizes the error of entering incorrect data concerning the operational information system of military forces; – performing initial identification of attacks specific to the given information system using a decision tree; – adapting to the type and duration of cyberattacks through multi-level adaptation of the artificial immune system; – conducting initial selection of antibodies for each swarm of the artificial immune system using an improved genetic algorithm; – training general-swarm antibodies using elite-swarm antibodies, thereby enabling deep learning; – replacing unfit individuals for search through antibody population renewal; – performing simultaneous solution search in multiple directions; – calculating the required amount of computational resources in cases where available resources are insufficient for the necessary calculations. An example application of the proposed method was conducted for cyberattack detection in an operational military force group. The results demonstrated an average increase in detection accuracy by 16 %, an average improvement in responsiveness by 12 %, and a high result convergence level of 95.23 %

This publication has 19 references indexed in Scilit: