Implicit Reward Strategies in Formations of New Rules in Reinforcement Learning

Student: Tsukanov Roman

Supervisor: Aleksandr I. Panov

Faculty: HSE Tikhonov Moscow Institute of Electronics and Mathematics (MIEM HSE)

Educational Programme: Applied Mathematics (Bachelor)

Final Grade: 9

Year of Graduation: 2018

Our main goal was to develop an algorithm of reinforcement learning which was able to implicitly differentiate agents by tasks. The solution does not use recieving explicit reward for doing required action by each agent. We also focus on main method of solving multi-agent systems by implementing common algorithms of single-agent reinforcement learning. In experimental part, the «ARM» enviroment we created which was used for modelling interaction of several manipulators able to move boxes in 2-dismentional space. We use «Actor-Critic» based on neural-network approximation of «Actor» and «Critic» for finding optimal behaviour of agents.

Full text (added May 25, 2018)

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