Swarming Mac Hall

CPSC 607 Project
John Brosz

Every day of the school year a swarm of students, staff, and unknowing outsiders descend upon MacEwen Student Center in search of food and seating. The goal of this project is to determine what aspects of this human crowd behavior can be modeled using simple agents. Success of this project may yield a useful tool for analyzing building usage patterns.

Up Project Proposal Current Work 607 Deliverables

Project Proposal

The Simulation

The first task in this project will be to create a model of the food court area of MacEwen Student Center (MSC). Figure 1 presents my rough model of the area of MSC I want to model.

Figure 1 - Rough, not-to-scale model of MSC. Green squares represent food locations; circles represent seating; red rectangles represent entries/exits.

Agents within the system will represent human individuals. These agents will have global knowledge of MSC (i.e., locations of restaurants and seating). However the agents will have only local knowledge of the location of other agents and occupation of seats.

My first goal in the system implementation is to have agents enter MSC at various entries, randomly walk, and then head toward an exit. In the second iteration, upon entering MSC the agents head towards a food establishment. Once an agent finds food, it will then head to a specific exit. In the third iteration, after obtaining food the agent will search for an unoccupied seat. I will also introduce controls for varying entry rate of the agents, time spent eating, and the time delay in picking up food.


The challenges of the project include handling collisions and introducing collision avoidance, creating a model of MSC, achieving reasonable behavior of agents in narrow aisles, use of a path finding techniques to move agents through the environment.

If the above challenges are met some additional areas I may explore are: use of pheromones to form queues for restaurants, incorporating groups of agents that travel and eat together, desire of agents to avoid crowded areas, agents that browse the food facilities, use of washrooms, and agents that travel at varying speeds.

Notes from Meeting

  • Make sure that I start simple, testing different behaviors individually before combining them.
  • Also start with simpler environment, rather than complex Mac Hall.
  • Emphasis is on why specific behaviours occur, rather than just presenting a mess.
  • Very open ended, exact details of what will be done at each still will depend heavily on result of previous stage(s).
  • There are definitely interesting possibilities in groups traveling together.
  • One idea that should make restaurant queues more managable is by having seperate 'entrances' and 'exits' from each restaurant area.


By John Brosz, January, 2006. Comments, suggestions or questions? If so, email me ([email protected]).