Risk-sensitive control of Markov decision processes
- In many revenue management applications risk-averse decision-making is crucial. In dynamic settings, however, it is challenging to find the right balance between maximizing expected rewards and minimizing various kinds of risk. In existing approaches utility functions, chance constraints, or (conditional) value at risk considerations are used to influence the distribution of rewards in a preferred way. Nevertheless, common techniques are not flexible enough and typically numerically complex. In our model, we exploit the fact that a distribution is characterized by its mean and higher moments. We present a multi-valued dynamic programming heuristic to compute risk-sensitive feedback policies that are able to directly control the moments of future rewards. Our approach is based on recursive formulations of higher moments and does not require an extension of the state space. Finally, we propose a self-tuning algorithm, which allows to identify feedback policies that approximate predetermined (risk-sensitive) target distributions. WeIn many revenue management applications risk-averse decision-making is crucial. In dynamic settings, however, it is challenging to find the right balance between maximizing expected rewards and minimizing various kinds of risk. In existing approaches utility functions, chance constraints, or (conditional) value at risk considerations are used to influence the distribution of rewards in a preferred way. Nevertheless, common techniques are not flexible enough and typically numerically complex. In our model, we exploit the fact that a distribution is characterized by its mean and higher moments. We present a multi-valued dynamic programming heuristic to compute risk-sensitive feedback policies that are able to directly control the moments of future rewards. Our approach is based on recursive formulations of higher moments and does not require an extension of the state space. Finally, we propose a self-tuning algorithm, which allows to identify feedback policies that approximate predetermined (risk-sensitive) target distributions. We illustrate the effectiveness and the flexibility of our approach for different dynamic pricing scenarios. (C) 2020 Elsevier Ltd. All rights reserved.…
Author details: | Rainer SchlosserORCiDGND |
---|---|
DOI: | https://doi.org/10.1016/j.cor.2020.104997 |
ISSN: | 0305-0548 |
Title of parent work (English): | Computers & operations research : and their applications to problems of world concern |
Subtitle (English): | a moment-based approach with target distributions |
Publisher: | Elsevier |
Place of publishing: | Oxford |
Publication type: | Article |
Language: | English |
Date of first publication: | 2020/06/02 |
Publication year: | 2020 |
Release date: | 2023/01/19 |
Tag: | Markov decision process; dynamic; dynamic programming; heuristics; pricing; risk aversion |
Volume: | 123 |
Article number: | 104997 |
Number of pages: | 14 |
Organizational units: | An-Institute / Hasso-Plattner-Institut für Digital Engineering gGmbH |
DDC classification: | 6 Technik, Medizin, angewandte Wissenschaften / 65 Management, Öffentlichkeitsarbeit / 650 Management und unterstützende Tätigkeiten |
Peer review: | Referiert |