Twin-system recurrent reinforcement learning for optimizing portfolio strategy

被引:1
作者
Park, Hyungjun [1 ,2 ]
Sim, Min Kyu [3 ]
Choi, Dong Gu [4 ]
机构
[1] Pohang Univ Sci & Technol, Dept Ind & Management Engn, 77 Cheongam Ro, Pohang 37673, Gyeongbuk, South Korea
[2] Pohang Univ Sci & Technol, Future IT Innovat Lab, 77 Cheongam Ro, Pohang 37673, Gyeongbuk, South Korea
[3] Seoul Natl Univ Sci & Technol, Dept Ind Engn, 232 Gongneung Ro, Seoul 01811, South Korea
[4] Pohang Univ Sci & Technol, Dept Ind & Management Engn, 77 Cheongam ro, Pohang 37673, Gyeongbuk, South Korea
基金
新加坡国家研究基金会;
关键词
Recurrent reinforcement learning; Twin-system approach; Constrained portfolio strategy; Mapping function; Markov decision process; OPTIMIZATION;
D O I
10.1016/j.eswa.2024.124193
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Portfolio management is important for sequential investment decisions in response to fluctuating financial markets. As portfolio management can be formulated as a sequential decision -making problem, it has been addressed using reinforcement learning in recent years. However, reinforcement learning methods face challenges in addressing portfolio management problems considering practical constraints. To overcome the limitations, this study proposes a twin -system approach, establishing a tractable twin that mirrors the original problem but with more manageable constraints and system dynamics. Once an optimized portfolio strategy is achieved within the tractable twin, the proposed mapping function translates it back to the original problem, ensuring the retention of optimized performance. Unlike the previous study, the proposed recurrent reinforcement learning method optimizes the portfolio strategy for a single agent managing all candidate assets. This method allows for comprehensive investment decisions by incorporating the features of candidate assets, leading to a more globally optimized portfolio strategy. Experimental studies demonstrate that the proposed method consistently outperforms benchmark strategies on the US sector and foreign exchange portfolios.
引用
收藏
页数:14
相关论文
共 40 条
  • [1] A constrained portfolio trading system using particle swarm algorithm and recurrent reinforcement learning
    Almahdi, Saud
    Yang, Steve Y.
    [J]. EXPERT SYSTEMS WITH APPLICATIONS, 2019, 130 : 145 - 156
  • [2] An adaptive portfolio trading system: A risk-return portfolio optimization using recurrent reinforcement learning with expected maximum drawdown
    Almahdi, Saud
    Yang, Steve Y.
    [J]. EXPERT SYSTEMS WITH APPLICATIONS, 2017, 87 : 267 - 279
  • [3] The mean-variance cardinality constrained portfolio optimization problem: An experimental evaluation of five multiobjective evolutionary algorithms
    Anagnostopoulos, K. P.
    Mamanis, G.
    [J]. EXPERT SYSTEMS WITH APPLICATIONS, 2011, 38 (11) : 14208 - 14217
  • [4] Bhatia P., 2018, Proceedings of the International Conference on Automated Planning and Scheduling, V29, P610, DOI DOI 10.1609/ICAPS.V29I1.3528
  • [5] Heuristics for cardinality constrained portfolio optimisation
    Chang, TJ
    Meade, N
    Beasley, JE
    Sharaiha, YM
    [J]. COMPUTERS & OPERATIONS RESEARCH, 2000, 27 (13) : 1271 - 1302
  • [6] Optimal Learning Algorithms for Stochastic Inventory Systems with Random Capacities
    Chen, Weidong
    Shi, Cong
    Duenyas, Izak
    [J]. PRODUCTION AND OPERATIONS MANAGEMENT, 2020, 29 (07) : 1624 - 1649
  • [7] Adversarial Attacks Against Reinforcement Learning-Based Portfolio Management Strategy
    Chen, Yu-Ying
    Chen, Chiao-Ting
    Sang, Chuan-Yun
    Yang, Yao-Chun
    Huang, Szu-Hao
    [J]. IEEE ACCESS, 2021, 9 : 50667 - 50685
  • [8] Conn A. R ..., 2000, Trust region methods
  • [9] Simulated annealing for complex portfolio selection problems
    Crama, Y
    Schyns, M
    [J]. EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2003, 150 (03) : 546 - 571
  • [10] Deep Direct Reinforcement Learning for Financial Signal Representation and Trading
    Deng, Yue
    Bao, Feng
    Kong, Youyong
    Ren, Zhiquan
    Dai, Qionghai
    [J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2017, 28 (03) : 653 - 664