Please use this identifier to cite or link to this item:
|Title:||RISK-AWARE PREFERENCE ELICITATION, ESTIMATION AND OPTIMIZATION||Authors:||HUANG WENJIE||Keywords:||Decision Making under Uncertainty, Risk Measure, Preference Elicitation, Markov Decision Processes, Stochastic Games, Reinforcement Learning||Issue Date:||25-Jan-2019||Citation:||HUANG WENJIE (2019-01-25). RISK-AWARE PREFERENCE ELICITATION, ESTIMATION AND OPTIMIZATION. ScholarBank@NUS Repository.||Abstract:||This thesis investigates realistic models for human risk preferences in decision-making under uncertainty, derives computationally tractable schemes, and develops theoretical analytic for the models and algorithms. In this thesis, we investigate three themes dealing with risk preferences. In this thesis, we successfully derive the decision theoretical and mathematical foundation of preference elicitation for quasi-concave choice functions (the widest class of choice functions representing diversification favoring behavior) on multi-attribute prospect space. Second, we propose a new class of risk measures that can be estimated accurately and efficiently by stochastic approximation technique and shows that many widely investigated risk measures follow its general mathematical representation. Finally, we propose powerful risk preference elicitation and estimation framework, and thus finally implement them in many classical and unified models in decision making under uncertainty, including expected utility theory, convex risk measures, satisficing and aspirational preference theory, Markov decision processes and stochastic games.||URI:||https://scholarbank.nus.edu.sg/handle/10635/156043|
|Appears in Collections:||Ph.D Theses (Open)|
Show full item record
Files in This Item:
|HuangWJ.pdf||2.75 MB||Adobe PDF|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.