Адаптивні системи автоматичного управління : міжвідомчий науково-технічний збірник. – 2023. – № 1 (42)

Постійне посилання зібрання

https://ela.kpi.ua/handle/123456789/55681

Переглянути

Перегляд Адаптивні системи автоматичного управління : міжвідомчий науково-технічний збірник. – 2023. – № 1 (42) за Автор "Albrekht, Y."

Зараз показуємо 1 - 2 з 2
  • Ескіз
    ДокументВідкритий доступ
    Learning rate in the reinforcement learning method for unknown location targets searching system
    (КПІ ім. Ігоря Сікорського, 2023) Albrekht, Y.; Pysarenko, A.
    The article explores the dependence of the system learning rate on the number of mutually independent modules in the reinforcement learning method. The study defines an environment with two types of objects that bring points to the final score and uses Deep Q Learning algorithms with 36 input data and 5 possible outcomes to conduct the experiment. The goal is to determine the optimal number of objects for which the use of reinforcement learning will give the best result for the same number of iterations. The research is part of a solution to the problem of creating a drone flock control system to find the position of objects in an unknown area.
  • Ескіз
    ДокументВідкритий доступ
    Unknown location targets searching system in known environment using reinforcement learning
    (КПІ ім. Ігоря Сікорського, 2023) Albrekht, Y.; Pysarenko, A.
    This article investigates two different approaches to searching for objects of a certain type in a known environment: with a centrally controlled system using individual modules that transmit information and by dividing the entire search area into smaller ones and using individual objects. The article conducts experiments using reinforcement learning algorithms to compare the learning speed and capabilities of a system with search modules and centralized control and a separate object to search for static objects with random locations in a known environment and to search for objects moving at a constant speed in a known environment. The article provides detailed information about the experimental design, including the definition of the parameters for reinforcement learning and the size of the input and output data for the neural network. The results of the experiments are presented graphically, demonstrating the effectiveness of reinforcement learning and the difference in the learning speed and capabilities of the two systems under study.