Grant: Discovery Grants Program - Individual

In distributed computing portions of a computational task are assigned to several processors. To accomplish the goal of such teamwork, some means permitting the exchange of information between the participating processors are usually in place. One possibility is to send messages between the stationary processors located at some nodes of the network. Another option is to admit processors moving along the edges of the network, or travelling inside some geometric environment. Such mobile processors (agents) may exchange the information when they meet (or they are sufficiently close to each other) or they leave the information at some places of the environment so that it is available to other agents visiting such places in the future. The subject of this research project is the study of algorithms for mobile agents working in networks as well as in the geometric environments. The main objectives of the project concern the feasibility of the fundamental problems of for mobile agents as well as the efficiency of their algorithmic solutions. The measures of efficiency involve algorithmic time complexity, memory and power used by the agents. The problems include collective exploration of the environment, rendezvous or gathering, pattern formation (moving of the agents in order to form a particular arrangement in the environment, e.g. to permit execution of some specific task). The applications of mobile agents are numerous. In robotics the mobile devices perform a task of terrain exploration or some target search. The software agents migrate over the network and their typical objective is remote data collection. Other applications come from wireless and ad-hoc networks, nanotechnology, medical computing, distributed spatial control, and many others. The results of this project will permit the deployment of more efficient teams of mobile agents in distributed computations, resulting in reducing network traffic and improved network reliability. The feasibility results will provide a better understanding of what is achievable in the field of mobile agents computations.