東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

FindBook

Google Book

Amazon

博客來

Reinforcement Learning for Stochastic Control and Games in Algorithmic Trading.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Reinforcement Learning for Stochastic Control and Games in Algorithmic Trading./
作者:	Ning, Xin.
面頁冊數:	1 online resource (121 pages)
附註:	Source: Dissertations Abstracts International, Volume: 84-09, Section: B.
Contained By:	Dissertations Abstracts International84-09B.
標題:	Statistics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30241688click for full text (PQDT)
ISBN:	9798377615750

Reinforcement Learning for Stochastic Control and Games in Algorithmic Trading.
Ning, Xin.

Reinforcement Learning for Stochastic Control and Games in Algorithmic Trading. - 1 online resource (121 pages)

Source: Dissertations Abstracts International, Volume: 84-09, Section: B.

Thesis (Ph.D.)--University of Toronto (Canada), 2023.

Includes bibliographical references

Algorithmic trading in electronic markets is a well-studied field in finance with a plethora of different possible approaches. This thesis explores how agents participating in electronic markets should optimally trade when accounting for latent factors and the behaviour of other participating agents. It investigates the problem by developing a modified Deep Q-Learning method for a single agent trader, generalizing the approach for multi-agent games using Nash equilibria, and constructing a new method for generating arbitrage-free implied volatility surfaces used for derivative pricing.The thesis contains three main parts. In the first part we take a model free approach to optimal trade execution and develop a variation of Deep Q-Learning to estimate the optimal actions of a trader. The model is a fully connected Neural Network trained using Experience Replay and Double DQN with input features given by the current state of the limit order book, other trading signals, and available execution actions, while the output is the Q-value function estimating the future rewards under an arbitrary action. We apply our model to nine different stocks and find that it significantly outperforms a baseline approach.In the second part, we generalize this approach into the multi-agent setting by developing a new data efficient Deep Q-learning methodology for model-free learning of Nash equilibria for general-sum stochastic games. The algorithm uses a local linear-quadratic approximation of the stochastic game, which leads to analytically solvable optimal actions. The approximation is parameterized by deep neural networks to provide sufficient flexibility to learn the environment without the need to experience all state-action pairs. We study symmetry properties of the algorithm stemming from label-invariant stochastic games and as a proof of concept, apply our algorithm to learning optimal trading strategies in various simulated competitive electronic markets.In the third part, we take the first steps in extending our approach into the derivatives market by proposing a new hybrid method for generating arbitrage-free implied volatility (IV) surfaces, which is a key feature used in derivative pricing. Our approach combines model-free Variational Autoencoders (VAEs) with continuous time stochastic differential equation (SDE) driven models. We focus on two classes of SDE models: regime switching models and Levy additive processes. By projecting historical surfaces onto the space of SDE model parameters, we obtain a distribution on the parameter subspace faithful to the data on which we then train a VAE. Arbitrage-free IV surfaces are then generated by sampling from the posterior distribution on the latent space, decoding to obtain SDE model parameters, and finally mapping those parameters to IV surfaces. We further refine the VAE model by including conditional features and demonstrate its superior generative out-of-sample performance.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9798377615750Subjects--Topical Terms:

517247
Statistics.
Subjects--Index Terms:

Algorithmic tradingIndex Terms--Genre/Form:

542853
Electronic books.

Reinforcement Learning for Stochastic Control and Games in Algorithmic Trading.
LDR:04459nmm a2200433K 4500 001 2359952
005 20230925131718.5
006 m o d
007 cr mn ---uuuuu
008 241011s2023 xx obm 000 0 eng d
020 $a 9798377615750
035 $a (MiAaPQ)AAI30241688
035 $a AAI30241688
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Ning, Xin. $3 1681096
245 1 0 $a Reinforcement Learning for Stochastic Control and Games in Algorithmic Trading.
264 0 $c 2023
300 $a 1 online resource (121 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 84-09, Section: B.
500 $a Advisor: Jaimungal, Sebastian.
502 $a Thesis (Ph.D.)--University of Toronto (Canada), 2023.
504 $a Includes bibliographical references
520 $a Algorithmic trading in electronic markets is a well-studied field in finance with a plethora of different possible approaches. This thesis explores how agents participating in electronic markets should optimally trade when accounting for latent factors and the behaviour of other participating agents. It investigates the problem by developing a modified Deep Q-Learning method for a single agent trader, generalizing the approach for multi-agent games using Nash equilibria, and constructing a new method for generating arbitrage-free implied volatility surfaces used for derivative pricing.The thesis contains three main parts. In the first part we take a model free approach to optimal trade execution and develop a variation of Deep Q-Learning to estimate the optimal actions of a trader. The model is a fully connected Neural Network trained using Experience Replay and Double DQN with input features given by the current state of the limit order book, other trading signals, and available execution actions, while the output is the Q-value function estimating the future rewards under an arbitrary action. We apply our model to nine different stocks and find that it significantly outperforms a baseline approach.In the second part, we generalize this approach into the multi-agent setting by developing a new data efficient Deep Q-learning methodology for model-free learning of Nash equilibria for general-sum stochastic games. The algorithm uses a local linear-quadratic approximation of the stochastic game, which leads to analytically solvable optimal actions. The approximation is parameterized by deep neural networks to provide sufficient flexibility to learn the environment without the need to experience all state-action pairs. We study symmetry properties of the algorithm stemming from label-invariant stochastic games and as a proof of concept, apply our algorithm to learning optimal trading strategies in various simulated competitive electronic markets.In the third part, we take the first steps in extending our approach into the derivatives market by proposing a new hybrid method for generating arbitrage-free implied volatility (IV) surfaces, which is a key feature used in derivative pricing. Our approach combines model-free Variational Autoencoders (VAEs) with continuous time stochastic differential equation (SDE) driven models. We focus on two classes of SDE models: regime switching models and Levy additive processes. By projecting historical surfaces onto the space of SDE model parameters, we obtain a distribution on the parameter subspace faithful to the data on which we then train a VAE. Arbitrage-free IV surfaces are then generated by sampling from the posterior distribution on the latent space, decoding to obtain SDE model parameters, and finally mapping those parameters to IV surfaces. We further refine the VAE model by including conditional features and demonstrate its superior generative out-of-sample performance.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Statistics. $3 517247
650 4 $a Computer science. $3 523869
650 4 $a Finance. $3 542899
653 $a Algorithmic trading
653 $a Implied volatility
653 $a Machine learning
653 $a Nash equilibrium
653 $a Reinforcement learning
653 $a Stochastic control
653 $a Electronic markets
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0463
690 $a 0984
690 $a 0508
690 $a 0800
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a University of Toronto (Canada). $b Statistics. $3 3175599
773 0 $t Dissertations Abstracts International $g 84-09B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30241688 $z click for full text (PQDT)