Using AI to Model Panic via NetLogo

Team: 50

School: Freedom High

Area of Science: Sociology

Interim:

Problem definition: Knowing how social interactions occur in a dangerous situation can help prevent some of the factors that cause people to make irrational panic decisions. During times of chaos, panic is what seems to make a group suffer. We intend to model how the panic spreads among the group. By determining what people as a whole will do in a bad situation we can determine how officials should treat a problem.

Project execution: Our project is to model patterns of reaction that occur in groups under pressure. Our team will do this by using artificial intelligence agents in a Netlogo base program. NetLogo is a program that involves agents that simulate a task through independent reaction to agent interactions. AI is a computer generated model that simulates decisions of an organism in which there is usually a set problem for AI to determine. We plan to program AI in Netlogo agents by gathering reactions made by human beings through television, newspapers, research papers, etc. these reactions will attempt to be modeled through AI decision makers. We will use multiple AI agents to react upon one agent catalyst. This will set the group of people panicking and fuel the thing causing our group to panic. There will be several different catalysts to manipulate the panic scenarios. These scenarios will include: outbreak of pandemic, an explosion of some kind, one agent threatening one agent in a group, one agent threatening the whole group, and news spreading panic throughout the system. The AI will make a randomized decision that is ‘illogical’ and will cause the rest of the catalysts to act. The program will set two main actions, the first being the scenario which we will change as required. The second will be the ‘lab rats’ or AI agent catalysts which we will manipulate to be where we want them during a specific time period. The AI agents will be instilled with decision making processes. We will then manipulate the scenario in order to investigate outputs in our panic research. By having many AI catalysts representing the people in a group panicking and by agent manipulators representing the panic scenarios, we can evaluate the situation in which groups panic.

Progress to Date: We have begun researching the complexity of group panic and how it can be applied in a simulated program. So far we have not found any existing programs that have been created to successfully model the sociology of group interactions during a series of panic situations. Through our research, we have discovered that in some instances, groups under pressure will not panic until the instant one person collapses under pressure causing a chain reaction. There are many complexities of the human mind we could not possibly comprehend or understand. But if we can get a grip on how humans react to threatening situations, as a unit, we may be able to simulate it.

Project projection: We predict that the sociology of a group will be determined by the reactions made by an individual catalyst and that people will follow the main decision of the first person who panics.

Citation:

1. By: Erik Auf der Heide: Common Misconceptions about Disasters: Panic, the “Disaster Syndrome,” and Looting.

2. By: Alexander Colhoun. On agent-based software engineering Nicholas R. Jennings Department of Electronics and Computer Science, University of Southampton, Southampton SO17 1BJ, UK. Received 21 September 1999, available online 31 July 2001.

http://goanna.cs.rmit.edu.au/~ssardina/courses/Roma07PhDcourse/lect10/Jennings_1999.pdf

3. By Raimo Tuomela: The philosophy of social practices: a collective acceptance view.

4. By Zili Zhang (Ph.D.), Chengqi Zhang: Agent-based hybrid intelligent systems: an agent-based framework for complex problem solving.

5. Kevin Kelly:

http://www.kk.org/outofcontrol/ch2-a.html

Team Members:

  William Barrett
  Chelsea Kibbee
  holly campbell
  Amy Ronan

Sponsoring Teacher: Richard Foust

Mail the entire Team