Showing 1–1 of 1 results for author: Lemburg, J P

Flexible forward improvement iteration for infinite time horizon Markovian optimal stopping problems

Authors: Sören Christensen, Albrecht Irle, Julian Peter Lemburg

Abstract: In this paper, we propose an extension of the forward improvement iteration algorithm, originally introduced in Irle (2006) and recently reconsidered in Miclo and Villeneuve (2021). The main new ingredient is a flexible window parameter describing the look-ahead distance in the improvement step. We consider the framework of a Markovian optimal stopping problem in discrete time with random discount… ▽ More In this paper, we propose an extension of the forward improvement iteration algorithm, originally introduced in Irle (2006) and recently reconsidered in Miclo and Villeneuve (2021). The main new ingredient is a flexible window parameter describing the look-ahead distance in the improvement step. We consider the framework of a Markovian optimal stopping problem in discrete time with random discounting and infinite time horizon. We prove convergence and show that the additional flexibility may significantly reduce the runtime. △ Less

Submitted 26 November, 2021; originally announced November 2021.

MSC Class: 60G40; 60J05; 91B28; 62L15