Temporal concatenation for Markov decision processes

被引:0
|
作者
Song, Ruiyang [1 ]
Xu, Kuang [2 ]
机构
[1] Stanford Univ, Dept Elect Engn, Stanford, CA 94305 USA
[2] Stanford Univ, Grad Sch Business, Stanford, CA USA
关键词
Markov decision process; Stochastic dynamic programming; HORIZON;
D O I
10.1017/S0269964821000206
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
We propose and analyze a temporal concatenation heuristic for solving large-scale finite-horizon Markov decision processes (MDP), which divides the MDP into smaller sub-problems along the time horizon and generates an overall solution by simply concatenating the optimal solutions from these sub-problems. As a "black box" architecture, temporal concatenation works with a wide range of existing MDP algorithms. Our main results characterize the regret of temporal concatenation compared to the optimal solution. We provide upper bounds for general MDP instances, as well as a family of MDP instances in which the upper bounds are shown to be tight. Together, our results demonstrate temporal concatenation's potential of substantial speed-up at the expense of some performance degradation.
引用
收藏
页码:999 / 1026
页数:28
相关论文
共 50 条
  • [41] Optimization of Markov decision processes under the variance criterion
    Xia, Li
    AUTOMATICA, 2016, 73 : 269 - 278
  • [42] Efficient Policies for Stationary Possibilistic Markov Decision Processes
    Ben Amor, Nahla
    El Khalfi, Zeineb
    Fargier, Helene
    Sabaddin, Regis
    SYMBOLIC AND QUANTITATIVE APPROACHES TO REASONING WITH UNCERTAINTY, ECSQARU 2017, 2017, 10369 : 306 - 317
  • [43] Concurrent Markov decision processes for robot team learning
    Girard, Justin
    Emami, M. Reza
    ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2015, 39 : 223 - 234
  • [44] An adaptation of particle swarm optimization for Markov decision processes
    Chang, HS
    2004 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN & CYBERNETICS, VOLS 1-7, 2004, : 1643 - 1648
  • [45] On optimality gaps for fuzzification in finite Markov decision processes
    Kageyama, Masayuki
    JOURNAL OF INTERDISCIPLINARY MATHEMATICS, 2008, 11 (01) : 77 - 88
  • [46] Singularly perturbed Markov decision processes in discrete time
    Liu, RH
    Zhang, Q
    Yin, G
    PROCEEDINGS OF THE 40TH IEEE CONFERENCE ON DECISION AND CONTROL, VOLS 1-5, 2001, : 2119 - 2124
  • [47] Aspects of Arranged Marriages and the Theory of Markov Decision Processes
    Amitrajeet a. Batabyal
    Theory and Decision, 1998, 45 : 241 - 253
  • [48] Interactive visualization for testing Markov Decision Processes: MDPVIS
    McGregor, Sean
    Buckingham, Hailey
    Dietterich, Thomas G.
    Houtman, Rachel
    Montgomery, Claire
    Metoyer, Ronald
    JOURNAL OF VISUAL LANGUAGES AND COMPUTING, 2017, 39 : 93 - 106
  • [49] Quantitative controller synthesis for consumption Markov decision processes
    Fu, Jianling
    Huang, Cheng-Chao
    Li, Yong
    Mei, Jingyi
    Xu, Ming
    Zhang, Lijun
    INFORMATION PROCESSING LETTERS, 2023, 180
  • [50] A primer on partially observable Markov decision processes (POMDPs)
    Chades, Iadine
    Pascal, Luz V.
    Nicol, Sam
    Fletcher, Cameron S.
    Ferrer-Mestres, Jonathan
    METHODS IN ECOLOGY AND EVOLUTION, 2021, 12 (11): : 2058 - 2072