Please use this identifier to cite or link to this item:
http://earchive.tpu.ru/handle/11683/80462
Title: | Обучение агентов в виртуальной среде KukaDiversObjectEnv |
Authors: | Залогин, Н. Е. |
metadata.dc.contributor.advisor: | Григорьев, Дмитрий Сергеевич |
Keywords: | reinforcement learning; robotic arm; PyBullet |
Issue Date: | 2024 |
Publisher: | Томский политехнический университет |
Citation: | Залогин, Н. Е. Обучение агентов в виртуальной среде KukaDiversObjectEnv / Н. Е. Залогин ; науч. рук. Д. С. Григорьев ; Национальный исследовательский Томский политехнический университет // Перспективы развития фундаментальных наук — Томск : Изд-во ТПУ, 2024. — Т. 7 : IT-технологии и электроника. — С. 42-44. |
Abstract: | The present study implements and compares the DQN, PPO, Parallel PPO, and Modified PPO algorithms in the PyBullet KukaDiverseObjectEnv environment. The algorithms are tested and evaluated in a simulated test mode to assess their performance. The experiments focus on metrics such as learning speed, stability, and task completion success rate. The results provide insights into the effectiveness of each algorithm in the tested environment, aiding in the optimization of reinforcement learning algorithms for complex environments and robotics |
URI: | http://earchive.tpu.ru/handle/11683/80462 |
Appears in Collections: | Материалы конференций |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
conference_tpu-2024-C21_V7_p42-44.pdf | 698,06 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.