ADAPTIVE ARTIFICIAL INTELLIGENCE SYSTEM FOR OPTIMIZING INFORMATION SYSTEMS

Gerasimov Vasily Aleksandrovich

The article proposes a new architecture of the artificial intelligence system aimed at solving the problems of optimization and modernization of information systems. The architecture allows optimization of various information systems using multi-criteria optimization methods, while remaining as flexible as possible. The approach allows artificial intelligence to automatically analyze and classify systems and propose optimization strategies. The main focus is on the application of multi-criteria optimization methods, self-learning and interaction with the user of the artificial intelligence system. Classical optimization methods and their applicability to the studied field of application and to the system as a whole are investigated. The problems of applying classical optimization methods in multi-criteria optimization are identified and a methodology for optimizing an artificial intelligence system based on two optimization methods is proposed: SLSQP and NSGA-II. The principle and necessity of the self-learning ability of such artificial intelligence systems are considered, and optimal methods of artificial intelligence training are proposed based on the study. Further studies of this system are considered in order to further improve it and increase efficiency in order to solve the problems of optimizing information systems.

Keywords:Artificial intelligence, neural network architecture, optimization, multi-criteria optimization methods.