Accession Number:

AD1058054

Title:

Some Stochastic Optimization Problems

Descriptive Note:

Technical Report,01 Apr 2014,30 Mar 2017

Corporate Author:

University of Southern California Los Angeles United States

Personal Author(s):

• Ross,Sheldon

Report Date:

2017-05-09

Pagination or Media Count:

17.0

Abstract:

We study a variety of stochastic optimization models related to assignment problems, scheduling models, queueing loss models, bandit problems, and others. We have obtained optimal policies in many cases, good heuristic ones in others, and have presented ways to analyze properties of these policies. In most cases, we have considered both static and dynamic policies. We have also analyzed stochastic systems related to queueing models with setup times, win probabilities in random knockout tournaments, and have obtained a new variation of the friendship paradox that your friends tend to have more friends than you do. In addition, we have made contributions to stochastic model theory by showing how to improve Poisson approximation bounds when applied to the probability that at least one of a set of events occurs. In all cases, we have presented a variety of possible applications of our model.

Descriptors:

• scheduling (production)
• stochastic processes
• computational modeling
• operations research
• optimization

Subject Categories:

• Statistics and Probability
• Operations Research
• Statistics and Probability
• Operations Research

Distribution Statement:

APPROVED FOR PUBLIC RELEASE