We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Probability

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

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 discounting and infinite time horizon. We prove convergence and show that the additional flexibility may significantly reduce the runtime.
Subjects: Probability (math.PR); Optimization and Control (math.OC); Computational Finance (q-fin.CP)
MSC classes: 60G40, 60J05, 91B28, 62L15
Cite as: arXiv:2111.13443 [math.PR]
  (or arXiv:2111.13443v1 [math.PR] for this version)

Submission history

From: Sören Christensen [view email]
[v1] Fri, 26 Nov 2021 11:54:44 GMT (65kb,D)

Link back to: arXiv, form interface, contact.