Sensitivity analysis of construction simulation
Abstract
In developing the CYCLONE methodology for simulating cyclic construction processes, a number of simplifying assumptions were made in order to provide a simple tool to be used by the construction personnel. It is the interest of this thesis to investigate if some of these simplifying assumptions have a significant effect in the results obtained from simulating a process using the MicroCYCLONE program. The three CYCLONE simplifying assumptions studied in this dissertation are: the simulation does not take into account the duration of a workday; an equipment breakdown is modeled using probabilistic branching, which is associated to an activity in the network where the equipment is being used; resources do not have attributes (e. g., equipment capacity, travel time as a function of capacity) associated to them. In order to investigate how these assumptions affect the simulation results, models are developed using the SLAM II methodology. SLAM II provides a simulation environment where each of the three simplifying assumptions can be removed from the models. The results obtained from the SLAM II models are compared against those obtained with the CYCLONE models. Two values in the results are looked at to provide a basis for comparison. They are the productivity of the process and the amount of time required to produce a given number of completed units. Two of the simplifying assumptions hold their validity when one takes into account that construction simulation results do not have to be extremely precise. These assumptions are: there is no need to take into account the workday duration in the simulation, and the resources can be modeled without associating any attributes to them. The assumption that equipment breakdowns can be modeled using probabilistic branching does not hold its validity. The results obtained when the equipment breakdown is modeled using a separate network, where the breakdown is associated with the equipment's own time distribution differ from the results predicted by MicroCYCLONE. The use of probabilistic branching is not an adequate way to model equipment breakdowns, unless a relationship is established between the equipment's utilization and the probability of the associated with the breakdown branch.
Degree
Ph.D.
Advisors
Halpin, Purdue University.
Subject Area
Civil engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.