Q LEARNING REGRESSION NEURAL NETWORK


Creative Commons License

SARIGUL M., AVCI M.

NEURAL NETWORK WORLD, vol.28, no.5, pp.415-431, 2018 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 28 Issue: 5
  • Publication Date: 2018
  • Doi Number: 10.14311/nnw.2018.28.023
  • Journal Name: NEURAL NETWORK WORLD
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.415-431
  • Keywords: reinforcement learning, q learning, q value function approximation, general regression neural network, kernel based regression, REINFORCEMENT, TEMPERATURE, MODEL
  • Çukurova University Affiliated: Yes

Abstract

In this work, a Nadaraya-Watson kernel based learning system which owns general regression neural network topology is adapted to Q learning method to evaluate a quick and efficient action selection policy for reinforcement learning problems. By means of the proposed method Q value function is generalized and learning speed of Q agent is accelerated. The training data of the developed neural network are obtained by a standard Q learning agent on closed-loop simulation system. The efficiency of the proposed method is tested on popular reinforcement learning benchmarks and its performance is compared with other popular regression methods and Q-learning utilized methods. QLRNN increased the learning performance and it learns faster than other methods on selected benchmarks. Test results showed the efficiency and the importance of the proposed network.