QUANTIFYING UNCERTAINTY IN SIMULATION MODELING

Abstract

Uncertainty can be defined as a state of either incomplete or otherwise bounded knowledge. Simulation models, and the engineering systems that they represent, often contain various types of uncertainty. Different approaches and theories can be applied to model these various types of uncertainty with a range of degrees in difficulty and accuracy. The objective of this paper is to explain the various types of uncertainty found in simulation models and to examine where uncertainty can be better represented or potentially reduced. To achieve this objective, a Monte Carlo Simulation model called the As-Planned Model is developed to estimate both cost and schedule using a risk-based approach for a simplified, Light Rail Transit construction project. After the project is complete, the As-Planned model is then compared to the project’s actual results. The resulting conclusions about various types of uncertainty are derived through both output comparison as well as uncertainty analysis.