AUTOMATION OF A PROJECT SOLUTION FOR SOFTWARE RECOVERY AFTER CYBERATTACKS

Abstract

This article examines the pressing scientific and practical problem of automating the recovery of software damaged by cyberattacks, an issue that is becoming particularly important given the growing number and complexity of cyber threats. An analysis of current scientific research was conducted, which showed that the main areas of development are the formalization of recovery processes based on mathematical models, as well as the application of artificial intelligence and machine learning methods. A representation of the process of recovering damaged software is proposed in the form of a set of interrelated tasks organized according to the principle of multilevel decomposition. Each task is described as a system of sets, including input data, constraints, solution options, evaluation criteria, models, and procedures for their implementation. This allows for the formalization of the decision-making process, the identification of dependencies between tasks, and the assurance of their algorithmic implementation. The paper also presents an example of a mathematical model for selecting the optimal method of software recovery following ransomware attacks. The model accounts for the structure of the software system, the set of defects, their properties and criticality, as well as available recovery methods, taking into account time, resource, and risk costs. An optimization objective function with weighting coefficients is proposed, which allows determining the most effective recovery strategy, taking into account the specified criteria. The results obtained demonstrate the feasibility of using formalized approaches