Presented to

Swarm computing involves devices operating like a swarm of insects. A swarm device is a small mobile robot with sensing capabilities. Therefore, swarm computing is significantly different from traditional computing methods in that swarm programs must be able to handle unknown environments and nodes that are capable of moving. Because swarm computing differs so drastically from traditional computing, traditional programming techniques will not work in swarm computing. Before such techniques can be developed for swarm computing, a better understanding needs to be gained of a swarm of computing devices. Through experimentation, I sought to gain an understanding of the basic principles of swarm computing. I used the libraries in the Swarm Development Group's package to create multiple programs implementing the same swarm behavior (primitives). Having wrote these programs, I ran them in the Swarm simulator and compare characteristics such as the power consumption of nodes. Having completed this project, I delivered a set of swarm primitives and information that I have gathered on basic swarm behavior. These primitives are not particularly useful by themselves but they can be combined and integrated to form swarm applications that actually perform a useful task. Ideally, other researchers can use my insights to develop future swarm programs. The researchers can then write programs for swarms of computational devices to accomplish such tasks as the exploration of Mars, clearing the blockage in arteries or other tasks that require the coordination of a multitude of devices. iii Table of Contents Abstract ...............................................................................................................................ii List of Tables................................