Creation of an Air Traffic Simulation Using
Agent Based Modelling
Sam Eberspacher
TJHSST Computer System Lab
2007-2008
April 4, 2008
Abstract
As the skies over the United States become increasingly crowded,
airports in the United States are increasingly stressed to adapt to this
increased demand. The goal of this project is to visually represent the
strain on airports and passengers as a variety of problems generate
record delays. By using agent based modelling, along with real air
traffic information, this simulation may accurately predict the prolif-
eration of delays through out airports in the United States.
1 Introduction
The purpose of this project is to visually represent the proliferation of a de-
lay throughout a system of airports. By using techiques such as agent based
modelling, the simulation will predict actual delays with decent accuracy.
Additionally by repeating the simulation multiple times, the simulation gen-
erates increasingly accurate results as the number of trials approached infin-
ity. While a simulation such as this would take a human enormous amounts
of time, a computer may be able to run a simulation of 24 hours in a matter
of minutes. Due to the scale of the problem, efficiency will be key for the
computer to run the simulation in a timely matter.
1
2 Development
2.1 Agent Based Modelling
In order to simulate such a large system, this project will use a technique
known as agent based modelling. The deveopment of a system using agent
based modelling is key for the success of the project. Each agent must interact
with other agents in the system in the most realistic way possible in order to
generate the most accurate results. One benefit of the agent based modelling
is that parameters for interaction between agents define the overall behaviour
of the system. This allows the programmer to work on much smaller problems
with the agent in order to alter the overall system.
2.2 Embedded Statistical Analysis
Embedded statistical analysis is a done when real time statistics are needed
in a si