BAYESIAN PROJECT MONITORING

DS 68-1: Proceedings of the 18th International Conference on Engineering Design (ICED 11), Impacting Society through Engineering Design, Vol. 1: Design Processes, Lyngby/Copenhagen, Denmark, 15.-19.08.2011

Year: 2011
Editor: Culley, S.J.; Hicks, B.J.; McAloone, T.C.; Howard, T.J. & Clarkson, P.J.
Author: Matthews, Peter C; Philip, Alex D M
Series: ICED
Section: Design Processes
Page(s): 69-78

Abstract

This paper studies how subtle signals that can be observed from the execution order of a project with several tasks can be used to diagnose potential problems that will hinder the project. Specifically, by representing the workflow of the project as a Markov Chain and observing how long the project takes to arrive at its first gateway, it is possible to infer the nature of any potential problems with the project. This diagnosis is achieved through using Bayesian methods, and provides a ranked list of potential problems, along with the probability for each problem. Two examples are used to illustrate how this approach works.

Keywords: UNCERTAINTY; PROJECT MANAGEMENT; MONTE CARLO SIMULATION; MARKOV CHAINS

Download