Risk-Sensitive Markov Decision Under Risk Constraints with Coherent Risk Measures

摘要

A Markov decision process with constraints of coherent risk measures is discussed. Risk-sensitive expected rewards under utility functions are approximated by weighted average value-at-risks, and risk constraints are described by coherent risk measures. In this paper, coherent risk measures are represented as weighted average value-at-risks with the best risk spectrum derived from decision maker's risk averse utility, and the risk spectrum can inherit the risk averse property of the decision maker's utility as weighting. To find risk levels for feasible ranges, firstly a risk-minimizing problem is discussed by mathematical programming. Next dynamic risk-sensitive reward maximization under risk constraints is investigated. Dynamic programming can not be applied to this dynamic optimization model, and we try other approaches. A few numerical examples are given to understand the obtained results.