Risk Aversion in Finite Markov Decision Processes Using Total Cost Criteria and Average Value at Risk

被引:0
|
作者
Carpin, Stefano [2 ]
Chow, Yin-Lam [1 ]
Pavone, Marco [1 ]
机构
[1] Stanford Univ, Stanford, CA 94305 USA
[2] Univ Calif Merced, Sch Engn, Merced, CA 95340 USA
关键词
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
In this paper we present an algorithm to compute risk averse policies in Markov Decision Processes (MDP) when the total cost criterion is used together with the average value at risk (AVaR) metric. Risk averse policies are needed when large deviations from the expected behavior may have detrimental effects, and conventional MDP algorithms usually ignore this aspect. We provide conditions for the structure of the underlying MDP ensuring that approximations for the exact problem can be derived and solved efficiently. Our findings are novel inasmuch as average value at risk has not previously been considered in association with the total cost criterion. Our method is demonstrated in a rapid deployment scenario, whereby a robot is tasked with the objective of reaching a target location within a temporal deadline where increased speed is associated with increased probability of failure. We demonstrate that the proposed algorithm not only produces a risk averse policy reducing the probability of exceeding the expected temporal deadline, but also provides the statistical distribution of costs, thus offering a valuable analysis tool.
引用
收藏
页码:335 / 342
页数:8
相关论文
共 50 条
  • [1] Markov Decision Processes with Average-Value-at-Risk criteria
    Baeuerle, Nicole
    Ott, Jonathan
    [J]. MATHEMATICAL METHODS OF OPERATIONS RESEARCH, 2011, 74 (03) : 361 - 379
  • [2] Markov Decision Processes with Average-Value-at-Risk criteria
    Nicole Bäuerle
    Jonathan Ott
    [J]. Mathematical Methods of Operations Research, 2011, 74 : 361 - 379
  • [3] AN OPTIMALITY SYSTEM FOR FINITE AVERAGE MARKOV DECISION CHAINS UNDER RISK-AVERSION
    Alanis-Duran, Alfredo
    Cavazos-Cadena, Rolando
    [J]. KYBERNETIKA, 2012, 48 (01) : 83 - 104
  • [4] Solutions of the average cost optimality equation for finite Markov decision chains: risk-sensitive and risk-neutral criteria
    Rolando Cavazos-Cadena
    [J]. Mathematical Methods of Operations Research, 2009, 70 : 541 - 566
  • [5] Solutions of the average cost optimality equation for finite Markov decision chains: risk-sensitive and risk-neutral criteria
    Cavazos-Cadena, Rolando
    [J]. MATHEMATICAL METHODS OF OPERATIONS RESEARCH, 2009, 70 (03) : 541 - 566
  • [6] Average criteria in denumerable semi-Markov decision chains under risk-aversion
    Rolando Cavazos-Cadena
    Hugo Cruz-Suárez
    Raúl Montes-De-Oca
    [J]. Discrete Event Dynamic Systems, 2023, 33 : 221 - 256
  • [7] Average criteria in denumerable semi-Markov decision chains under risk-aversion
    Cavazos-Cadena, Rolando
    Cruz-Suarez, Hugo
    Montes-De-Oca, Raul
    [J]. DISCRETE EVENT DYNAMIC SYSTEMS-THEORY AND APPLICATIONS, 2023, 33 (03): : 221 - 256
  • [8] An average-value-at-risk criterion for Markov decision processes with unbounded costs
    Liu, Qiuli
    Ching, Wai-Ki
    Zhang, Junyu
    Wang, Hongchu
    [J]. FRONTIERS OF MATHEMATICS IN CHINA, 2022, 17 (04) : 673 - 687
  • [9] An average-value-at-risk criterion for Markov decision processes with unbounded costs
    Qiuli Liu
    Wai-Ki Ching
    Junyu Zhang
    Hongchu Wang
    [J]. Frontiers of Mathematics in China, 2022, 17 : 673 - 687
  • [10] MINIMUM AVERAGE VALUE-AT-RISK FOR FINITE HORIZON SEMI-MARKOV DECISION PROCESSES IN CONTINUOUS TIME
    Huang, Yonghui
    Guo, Xianping
    [J]. SIAM JOURNAL ON OPTIMIZATION, 2016, 26 (01) : 1 - 28