Decision-Theoretic Methods in Simulation Optimization

Abstract

Simulation optimization is the process of selecting the best among a collection of options, where each option can only be evaluated through stochastic simulation. A good simulation optimization algorithm enables finding the best option, or a very good option, in a reasonable amount of time. Because time is limited, the quality of the simulation optimization algorithm used can often determine whether we are able to discover a high quality option, or merely a mediocre one. The Department of Defense uses stochastic simulators to make a variety of critical strategic decisions, and good simulation optimization algorithms are critical to this decision-making process. The work in this project uses decision-theory to study and improve the decisions made by simulation optimization algorithms, providing algorithms for a number of distinct simulation optimization problems that improve performance over previous state-of-the-art algorithms.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 24, 2014
Accession Number
ADA610908

Entities

People

  • Peter Frazier

Organizations

  • Cornell University College of Engineering

Tags

Communities of Interest

  • Biomedical
  • Human Systems

DTIC Thesaurus Topics

  • Air Force Research Laboratories
  • Algorithms
  • Computational Science
  • Computer Programming
  • Computers
  • Data Mining
  • Dynamic Programming
  • Engineering
  • Engineers
  • Industrial Engineering
  • Information Science
  • Machine Learning
  • Materials Science
  • Operations Research
  • Probability
  • Students
  • Systems Engineering

Fields of Study

  • Computer science

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Defense Acquisition Program Management
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.