PROBABILISTIC SIMULATION STUDIES FOR REPETITIVE CONSTRUCTION PROCESSES

Before full adaptation of simulation techniques for the analysis and design of construction operations can be implemented, the appropriate statistical tools must be understood and applied as part of the simulation experiment. Most simulation models in construction can be treated as stochastic models. The proper analysis of such models requires: (1) Application of input modeling techniques; (2) appropriate analysis of output parameters of concern based on multiple runs; and (3) validation and verification of the results. This paper describes techniques applicable to the simulation of repetitive construction operations and presents a practical example application to demonstrate how they can be applied. Procedures for selecting input models, methods for solving for the parameters of selected distributions, and goodness-of-fit testing for construction data are reviewed. The discussions on output analysis were limited to simulation output that is normal because it is frequently encountered in construction simulation. Methods for checking the normality of output data, building confidence intervals for various output parameters, and validation of simulation models are also addressed. © ASCE.