Risk Analysis in Construction

The work reported in this thesis centres around the development of a risk analysis computer software package which operates in conjunction with PERTMASTER ADVANCE, a widely-used commercial project planning software package. The risk analysis technique used herein is the so- called 'Monte Carlo simulation technique' whereby a number of project simulations, each of which makes use of a set of project data values (such as task durations) chosen randomly from within specified probability distributions, is carried out with a view to producing a particular solution (such as the project duration) defined by a single probability distribution. In such a way, the degree of uncertainty associated with the particular solution, which results from the individual degrees of uncertainty associated with each of the data values on which such solution is based, can be determined. The software package comprises the following programs: PERTRA1 - an interactive program which abstracts specific project data from a datafile generated by PERTMASTER ADVANCE. PERTRA2 - an interactive program which facilitates the keyboard input of task duration probability distribution data. PERTRA3 - a 'number- crunching' program which carries out the risk analysis on the basis of data generated by PERTRA1 and PERTRA2. The results of the analysis are written to a user- specified datafile in such a form as to enable subsequent graphical presentation thereof by widely available commercial 'graphics' software packages. All three programs are written in FORTRAN 77 and can be run on any IBM-compatible personal computer using the MS-DOS version 3.3 or higher