A note on a new class of recursive utilities in Markov decision processes
Tom 44 / 2017
Applicationes Mathematicae 44 (2017), 149-161
MSC: Primary 90C40; Secondary 90C39, 91B70.
DOI: 10.4064/am2317-1-2017
Opublikowany online: 24 May 2017
Streszczenie
This paper deals with Markov decision processes on a general state space under standard compactness-continuity assumptions. The purpose is to obtain a new class of so-called recursive utilities with the aid of the entropic risk measure. Within this framework we show that there exists a stationary policy for a discounted payoff problem in the infinite time horizon. Our result is illustrated by examples.
