Full-information best choice game with hint (Q2009175)

The author considers a modification of the classical full-information best choice problem in the context of Markov decision processes where a decision maker aims to select the best opportunity based only on the exact values of the observed sequence. The modification is adding a second player who can either propose additional information or block the observed object and demand an extortion. An optimal reward for the second player and the best moment to interrupt the decision process has been obtained and the situation when the number of observations tends to infinity has been studied.