In this paper the use of rank correlations in simulating construction costs is investigated. One suggested methodology for generating correlated random numbers using rank correlations is reviewed and compared with traditional methods based on Pearson correlations. This methodology is the basis for the design of several simulation software packages commonly used by analysts and estimators. Because of this it is important to evaluate the effectiveness of this approach in probabilistic analysis of construction costs. A set of real-life construction costs is used to test the effectiveness of the suggested methodology in simulating the distribution of costs. Several tests of hypotheses are executed to compare the distribution of simulated data with actual data. It is shown that rank correlations can model data dependency as effectively as Pearson correlations on this data set. Suggestions are made regarding future work in this area.
[1]
S. Siegel,et al.
Nonparametric Statistics for the Behavioral Sciences
,
2022,
The SAGE Encyclopedia of Research Design.
[2]
G. S. Fishman.
Principles of Discrete Event Simulation
,
1978
.
[3]
R. Iman,et al.
A distribution-free approach to inducing rank correlation among input variables
,
1982
.
[4]
R. Iman,et al.
Rank correlation plots for use with correlated input variables
,
1982
.
[5]
James E. Diekmann.
Probabilistic Estimating: Mathematics and Applications
,
1983
.
[6]
Simaan M. AbouRizk,et al.
PROBABILISTIC SIMULATION STUDIES FOR REPETITIVE CONSTRUCTION PROCESSES
,
1990
.
[7]
Ali Touran,et al.
Monte Carlo Technique with Correlated Random Variables
,
1992
.