Lots of Simple, Little Things Interacting in Cooperative, Intelligent Ways: What Can We Learn From a Swarm of Ants - A Tutorial

Abstract:

We have developed a Flash-based, web-deliverable, learning tool for understanding Emergent Behavior concepts. In many areas of science, business, and society, we find leaderless processes exhibit not only intelligence, but self-organization. While these processes are flexible, relatively inexpensive, and amazing to watch, they are often counter-intuitive. We built this tool to help people understand the concepts, trade-offs and implementation mechanisms of these increasingly important design principle. We propose a tutorial on our Emergent Behavior Simulation Tool (EBST). The tool is designed to help individuals who are not familiar with swarms and emergent behavior to become conversant with the concepts, obstacles and processes of an emergent system. EBST makes emergent systems and bottom-up design quickly accessible to interested neophytes. The contextual scenarios are built by swarm researchers, but provide an interface accessible to all. The tool provides an environment in which the consequences of individual actions can be observed at the global level, and then quickly changed based on observations. Yet the scenarios are specific, and goal-oriented, so that the swarm-layperson is not left to fend for themselves in a sea of syntax, and open-ended possibilities.

Our tutorial at ASCUE will lead participants through three scenarios with increasing fidelity and consideration of additional factors. Each simulation presents an objective, a set of constraints, and a swarm of programmable individuals. In the first simulation, called “Ants and Bridges”, the goal is to have ants traverse a water obstacle by building a bridge with their bodies in order to get to the other side. The user “programs” the ants’ behavior by constructing a Finite State Automata and by setting priorities on the ant’s sensors. The second simulation, called “Ants, Bridge, and Food” is an extension in which the ants need to replenish their energy to survive. The third simulation, called “Mine Sweeping”, has a group of robots clear a shoreline landing area of mines. The mine-sweeping simulation has many more states and sensors to work with which increases the complexity of “programming the swarm”.