Analysis of Beta Distribution for Subjective Uncertainty in Cost Models

Abstract

Subjective uncertainty exists within the realm of cost estimation. Typical methodology for subjective uncertainty involves elicitation from a subject matter expert to provide a high, low, and most likely value - defining a triangular distribution - to model said uncertainty. This manuscript explores ways to leverage research on elicitation geared towards defining a triangular distribution and provide a simple conversion to a beta distribution usable by cost analysts with various degrees of mathematical knowledge. Furthermore, this manuscript attempts to demonstrate the benefits of using a beta distribution through its application as a conjugate prior for Bayesian updating in cost models.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 26, 2020
Accession Number
AD1101752

Entities

People

  • Ryan D Stafford

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Energy and Power Technologies
  • Human Systems

DTIC Thesaurus Topics

  • Air Force
  • Bayesian Inference
  • Bayesian Networks
  • Best Practices
  • Computational Science
  • Cost Estimates
  • Delphi Method
  • Department Of Defense
  • Information Science
  • Knowledge Management
  • Monte Carlo Method
  • Operations Research
  • Probability
  • Probability Distributions
  • Risk Analysis
  • Software Development
  • Statistical Analysis

Readers

  • Computational Modeling and Simulation
  • Fluid Dynamics.
  • Life Cycle Cost Analysis

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - DoD AI Strategy