Estimating parameters of the triangular distribution using nonstandard information

  • Authors:
  • Seratun Jannat;Allen G. Greenwood

  • Affiliations:
  • Mississippi State University, Mississippi State, MS;Mississippi State University, Mississippi State, MS

  • Venue:
  • Proceedings of the Winter Simulation Conference
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

The triangular distribution is commonly used in simulation projects to represent probabilistic processes in absence of detailed data. The distribution can take on a variety of shapes and requires three easy to estimate basic parameters -- minimum, maximum, and most likely. This paper considers two situations where different information is available than the three basic parameters. The paper provides means to use different information to estimate the remaining distribution parameters. The first situation commonly occurs in practice. For example, detail data may not be available, but the mean is known; thus, only two basic parameters need to be specified. The second situation occurs in research where controlled comparisons need to be made. For example, in order to understand the effect of variability on a system, means and general shape need to be held constant; thus, by fixing these two characteristics, only one of the basic parameters needs to be specified.