Open AccessChoice of cargo delivery option in multimodal connection based on reinforcement learning
Author(s) -
А П Бадецкий,
Oksana Medved'
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2131/3/032103
Subject(s) - reinforcement learning , connection (principal bundle) , computer science , convergence (economics) , artificial intelligence , basis (linear algebra) , machine learning , mathematical optimization , engineering , mathematics , economics , economic growth , geometry , structural engineering
The article discusses the issues of choosing a route and an option of cargo flows in multimodal connection in modern conditions. Taking into account active development of artificial intelligence and digital technologies in all types of production activities, it is proposed to use reinforcement learning algorithms to solve the problem. An analysis of the existing algorithms has been carried out, on the basis of which it was found that when choosing a route option for cargo in a multimodal connection, it would be useful to have a qualitative assessment of terminal states. To obtain such an estimate, the Q-learning algorithm was applied in the article, which showed sufficient convergence and efficiency.