• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Optimization Methods in Reinforcement Learning

Student: Haoyang Xu

Supervisor: Alexander Sirotkin

Faculty: St. Petersburg School of Physics, Mathematics, and Computer Science

Educational Programme: Big Data Analysis for Business, Economy, and Society (Master)

Year of Graduation: 2020

With continuous research on reinforcement learning, more and more types of reinforcement learning algorithms are emerging, tasks are becoming more and more complex, and it is becoming increasingly difficult to tune parameters. There are more and more algorithms for reinforcement learning, performance is becoming more and more complex, and the algorithm is more complex. There are more and more problems, setting up the algorithm is becoming more and more complicated, and the applicability of the algorithm is also unstable. Therefore, the parameter optimization algorithm has become an important task. The applicability of the algorithm to different environments is also very different. This article uses three classic reinforcement learning algorithms with DQN, DDPG, PPO in cartpole-v0 and Pendulum-v0 environments. Ten optimization methods apply to these three types of reinforcement training in two gaming environments. Summarize the methods of reinforcement learning optimization methods, compare the effectiveness of these ten optimization methods, and optimize more efficient algorithms based on basic algorithms.

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses