Scholarlink

Abstract <p>The possibility of integrating fuzzy logic and reinforcement learning methods is considered. An&#xa0;implementation of a self-organizing fuzzy neural Q-network using the fuzzy conservative Q-learning algorithm is presented. The process of automatic tuning of fuzzy Q-network is described. The results of testing the fuzzy conservative Q-learning algorithm and the fuzzy Q-learning algorithm on two model problems, CartPole and MountainCar, are presented. Their comparative analysis is considered when solving these problems. In the conclusions, the main advantages of the self-organizing fuzzy Q-network are described in detail. The work was carried out in the context of research and development of mathematical and software support for intelligent real-time decision support systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine Learning Based on a Self-Organizing Fuzzy Neural Network and Reinforcement Learning

  • 期刊论文

  • 发表日期：

    2025年12月1日

查看全文
  • Alexander Pavlovich Eremeev,
  • Maxim Dmitrievich Sergeev,
  • Vladimir Sergeevich Petrov

摘要

Abstract

The possibility of integrating fuzzy logic and reinforcement learning methods is considered. An implementation of a self-organizing fuzzy neural Q-network using the fuzzy conservative Q-learning algorithm is presented. The process of automatic tuning of fuzzy Q-network is described. The results of testing the fuzzy conservative Q-learning algorithm and the fuzzy Q-learning algorithm on two model problems, CartPole and MountainCar, are presented. Their comparative analysis is considered when solving these problems. In the conclusions, the main advantages of the self-organizing fuzzy Q-network are described in detail. The work was carried out in the context of research and development of mathematical and software support for intelligent real-time decision support systems.