Logconcave reward functions and optimal stopping rules of threshold form

Shoou-Ren Hsiau, Yi Shen Lin, Yi Ching Yao

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

In the literature, the problem of maximizing the expected discounted reward over all stopping rules has been explicitly solved for a number of reward functions (including (max{x, 0})v, v > 0, in particular) when the underlying process is either a random walk in discrete time or a Levy process in continuous time. All of such reward functions are increasing and logconcave while the corresponding optimal stopping rules have the threshold form. In this paper, we explore the close connection between increasing and logconcave reward functions and optimal stopping rules of threshold form. In the discrete case, we show that if a reward function defined on ℤ is nonnegative, increasing and logconcave, then the optimal stopping rule is ofthreshold form provided the underlying random walk is skip-free to the right. In the continuous case, it is shown that for a reward function defined on ℝ which is nonnegative, increasing, logconcave and right-continuous, the optimal stopping rule is of threshold form provided the underlying process is a spectrally negative lévy process. Furthermore, we also establish the necessity of logconcavity and monotonicity of a reward function in order for the optimal stopping rule to be of threshold form in the discrete (continuous, resp.) case when the underlying process belongs to the class of Bernoulli random walks (Brownian motions, resp.) with a downward drift. These results together provide a partial characterization of the threshold structure of optimal stopping rules.

Original languageEnglish
Article number120
JournalElectronic Journal of Probability
Volume19
DOIs
Publication statusPublished - 2014 Dec 27

Fingerprint

Optimal Stopping Rule
Log-concave
Reward
Random walk
Non-negative
Log-concavity
Stopping Rule
Lévy Process
Form
Stopping rule
Optimal stopping
Bernoulli
Brownian motion
Monotonicity
Continuous Time
Discrete-time
Partial

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

@article{bb41c5bfac0e4d13b825d2744a19fe8d,
title = "Logconcave reward functions and optimal stopping rules of threshold form",
abstract = "In the literature, the problem of maximizing the expected discounted reward over all stopping rules has been explicitly solved for a number of reward functions (including (max{x, 0})v, v > 0, in particular) when the underlying process is either a random walk in discrete time or a Levy process in continuous time. All of such reward functions are increasing and logconcave while the corresponding optimal stopping rules have the threshold form. In this paper, we explore the close connection between increasing and logconcave reward functions and optimal stopping rules of threshold form. In the discrete case, we show that if a reward function defined on ℤ is nonnegative, increasing and logconcave, then the optimal stopping rule is ofthreshold form provided the underlying random walk is skip-free to the right. In the continuous case, it is shown that for a reward function defined on ℝ which is nonnegative, increasing, logconcave and right-continuous, the optimal stopping rule is of threshold form provided the underlying process is a spectrally negative l{\'e}vy process. Furthermore, we also establish the necessity of logconcavity and monotonicity of a reward function in order for the optimal stopping rule to be of threshold form in the discrete (continuous, resp.) case when the underlying process belongs to the class of Bernoulli random walks (Brownian motions, resp.) with a downward drift. These results together provide a partial characterization of the threshold structure of optimal stopping rules.",
author = "Shoou-Ren Hsiau and Lin, {Yi Shen} and Yao, {Yi Ching}",
year = "2014",
month = "12",
day = "27",
doi = "10.1214/EJP.v19-3745",
language = "English",
volume = "19",
journal = "Electronic Journal of Probability",
issn = "1083-6489",
publisher = "Institute of Mathematical Statistics",

}

Logconcave reward functions and optimal stopping rules of threshold form. / Hsiau, Shoou-Ren; Lin, Yi Shen; Yao, Yi Ching.

In: Electronic Journal of Probability, Vol. 19, 120, 27.12.2014.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Logconcave reward functions and optimal stopping rules of threshold form

AU - Hsiau, Shoou-Ren

AU - Lin, Yi Shen

AU - Yao, Yi Ching

PY - 2014/12/27

Y1 - 2014/12/27

N2 - In the literature, the problem of maximizing the expected discounted reward over all stopping rules has been explicitly solved for a number of reward functions (including (max{x, 0})v, v > 0, in particular) when the underlying process is either a random walk in discrete time or a Levy process in continuous time. All of such reward functions are increasing and logconcave while the corresponding optimal stopping rules have the threshold form. In this paper, we explore the close connection between increasing and logconcave reward functions and optimal stopping rules of threshold form. In the discrete case, we show that if a reward function defined on ℤ is nonnegative, increasing and logconcave, then the optimal stopping rule is ofthreshold form provided the underlying random walk is skip-free to the right. In the continuous case, it is shown that for a reward function defined on ℝ which is nonnegative, increasing, logconcave and right-continuous, the optimal stopping rule is of threshold form provided the underlying process is a spectrally negative lévy process. Furthermore, we also establish the necessity of logconcavity and monotonicity of a reward function in order for the optimal stopping rule to be of threshold form in the discrete (continuous, resp.) case when the underlying process belongs to the class of Bernoulli random walks (Brownian motions, resp.) with a downward drift. These results together provide a partial characterization of the threshold structure of optimal stopping rules.

AB - In the literature, the problem of maximizing the expected discounted reward over all stopping rules has been explicitly solved for a number of reward functions (including (max{x, 0})v, v > 0, in particular) when the underlying process is either a random walk in discrete time or a Levy process in continuous time. All of such reward functions are increasing and logconcave while the corresponding optimal stopping rules have the threshold form. In this paper, we explore the close connection between increasing and logconcave reward functions and optimal stopping rules of threshold form. In the discrete case, we show that if a reward function defined on ℤ is nonnegative, increasing and logconcave, then the optimal stopping rule is ofthreshold form provided the underlying random walk is skip-free to the right. In the continuous case, it is shown that for a reward function defined on ℝ which is nonnegative, increasing, logconcave and right-continuous, the optimal stopping rule is of threshold form provided the underlying process is a spectrally negative lévy process. Furthermore, we also establish the necessity of logconcavity and monotonicity of a reward function in order for the optimal stopping rule to be of threshold form in the discrete (continuous, resp.) case when the underlying process belongs to the class of Bernoulli random walks (Brownian motions, resp.) with a downward drift. These results together provide a partial characterization of the threshold structure of optimal stopping rules.

UR - http://www.scopus.com/inward/record.url?scp=84942239478&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84942239478&partnerID=8YFLogxK

U2 - 10.1214/EJP.v19-3745

DO - 10.1214/EJP.v19-3745

M3 - Article

AN - SCOPUS:84942239478

VL - 19

JO - Electronic Journal of Probability

JF - Electronic Journal of Probability

SN - 1083-6489

M1 - 120

ER -