В статье рассматриваются проблемы разработки и исследования нейросетевых и гибридных алгоритмов обучения для повышения эффективности функционирования интеллектуальных систем с целью поддержки принятия решений в сложных средах. The article is devoted to the problems for solving difficult problems, such as prediction, planning, pattern recognition and knowledge discovery in a number of application areas: bioinformatics, speech and language, image and video analysis, other engineering disciplines. Most of these publications deal with static process, assuming that the process is represented adequately by the data available at present and that it does not change over time.