Statistical PERT: An Improved Subnetwork Analysis Procedure

Abstract

Statistical PERT is a new procedure for obtaining information about the distribution of a project's completion time when the project is comprised of a large number of activities and the time required to complete an individual activity once it can be begun is a random variable. The project is represented as an acyclic network whose arcs correspond to the project activities. This network is simplified by replacing various activity configurations by single equivalent activities and then decomposed into several subnetworks. The distribution and moments of each subnetwork's completion time are bounded and approximated on the basis of two points from each activity's completion time distribution by using some mathematical programming techniques and a new result in the theory of networks. The project's completion time distribution is then approximated by combining the approximate subnetwork distributions. This report describes several refinements in the subnetwork analysis procedure.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 1975
Accession Number
ADA020470

Entities

People

  • Herman Otto Hartley
  • R. K. Spoeri
  • R. L. Sielken Jr.

Organizations

  • Texas A&M University

Tags

Communities of Interest

  • Advanced Electronics
  • Energy and Power Technologies
  • Ground and Sea Platforms
  • Weapons Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Applied Mathematics
  • Computer Programming
  • Computer Programs
  • Computers
  • Distribution Functions
  • Mathematical Programming
  • Mathematics
  • Monte Carlo Method
  • Navy
  • New York
  • Operations Research
  • Probability
  • Random Variables
  • Simplex Method
  • Statistical Samples
  • Statistics

Fields of Study

  • Computer science

Readers

  • Applied Combinatorial Optimization and Logic Circuit Design.
  • Mathematical Modeling and Probability Theory.
  • Technical Research and Report Writing.